Contents at a Glance

Introduction
1 An Overview of SOAP 11
2 SOAP in Theory 33
3 An Overview of Security Issues for SOAP 61
4 Using SOAP to Create a Simple Application 81
5 Migrating an Application from DCOM to SOAP 121
6 Creating Remote Access Utilities 161
7 Creating Data Entry Forms and Surveys 195
8 Providing Remote Database Access 237
9 Moving to Web-based Applications 283
10 Working with PDAs 321

Using A
B
C
Appendixes
SOAP Data Types and Data Type Conversions 345
Microsoft Biztalk and SOAP 355
Third-Party Tool Reference 371

SOAP
D SOAP for Visual C++ Developers 389

Glossary 407

Index 427

John Paul Mueller

201 W. 103rd Street

Indianapolis, Indiana 46290
Special Edition Using SOAP Associate Publisher
Dean Miller
Copyright  2002 by Que
Executive Editor
All rights reserved. No part of this book shall be repro- Candace Hall
duced, stored in a retrieval system, or transmitted by any Acquisitions Editor
means, electronic, mechanical, photocopying, recording, Michelle Newcomb
or otherwise, without written permission from the pub- Development Editors
lisher. No patent liability is assumed with respect to the Sarah Robbins
use of the information contained herein. Although every Maureen McDaniel
precaution has been taken in the preparation of this book, Managing Editor
the publisher and author assume no responsibility for Thomas F. Hayes
errors or omissions. Nor is any liability assumed for dam-
Project Editor
ages resulting from the use of the information contained Karen S. Shields
herein.
Copy Editors
International Standard Book Number: 0-7897-2566-5 Karen Gill
Kay Hoskin
Library of Congress Catalog Card Number: 2001090465
Indexer
Printed in the United States of America Johnna Dinse

First Printing: September, 2001 Proofreader

Marcia Deboy
04 03 02 01 4 3 2 1
Technical Editor
Trademarks Russ Mullen
Team Coordinator
All terms mentioned in this book that are known to be
Cindy Teeters
trademarks or service marks have been appropriately capi-
talized. Que cannot attest to the accuracy of this informa- Media Developer
Michael Hunter
tion. Use of a term in this book should not be regarded as
affecting the validity of any trademark or service mark. Interior Designer
Ruth Harvey
Warning and Disclaimer
Cover Designers
Every effort has been made to make this book as complete Dan Armstrong
Ruth Harvey
and as accurate as possible, but no warranty or fitness is
implied. The information provided is on an “as is” basis. Page Layout
The author and the publisher shall have neither liability Tricia Bronkella
Rebecca Harmon
nor responsibility to any person or entity with respect to
Cheryl Lynch
any loss or damages arising from the information con-
tained in this book.
Contents
Introduction 1 Working with the Web Service
Description Language (WSDL) 48
1 An Overview of SOAP 11 Understanding Universal Discovery,
Description, and Integration
What Is SOAP? 13 (UDDI) 49
How SOAP Differs from DCOM and Putting Everything Together 50
CORBA 15
DCOM Wire Protocol 15 Using SOAP to Move Data 52
CORBA IIOP 19 Current SOAP Implementation
Java RMI 20 Problems 53
SOAP and XML-RPC 56
SOAP, HTTP, and XML 22
SOAP and XML Protocol
Problems Solved by Using SOAP 24 (XMLP) 56

Performance Issues 26 Understanding SOAP Attachments 57

Tradeoffs of Using SOAP 26
Case Study 58
When SOAP Is Faster 27

SOAP and the Web Server 28 3 An Overview of Security Issues

for SOAP 61
Why Make the Move to SOAP? 28
Introduction 62
Case Study 29
Definition of a Problem 29 Understanding SOAP Privacy and
Light at the End of the Tunnel 30 Security Issues 63
Perfection, an Ongoing Process 30 What Are the Issues? 64
User Privacy Issues 65
2 SOAP in Theory 33 Data Security Issues 66

Dissecting the SOAP Message 35 Security Standards You Should Know

About 67
Viewing the HTTP Portion of
HTTP Authentication Framework 71
SOAP 37
Secure/Multipurpose Internet Mail
Viewing the XML Portion of
Extensions (S/MIME) 73
SOAP 38
Secure Socket Layer (SSL) 74
Working with the SOAP Message 39
User Identification Issues 75
Using SOAP with Your Current
Code 45 Using Smart Cards 75
Using Biometrics 76
Discovering SOAP Services 47
Understanding Discovery of Web Where Do You Go from Here? 77
Services (DISCO) 48 Case Study 78
iv Special Edition Using SOAP

4 Using SOAP to Create a Simple 5 Migrating an Application from DCOM

Application 81 to SOAP 121

Introduction 82 Introduction 122

An Overview of Microsoft’s SOAP SOAP Application Conversion

Toolkit 83 Prerequisites 123
Toolkit Features 84 An Overview of the Conversion
Three Types of Microsoft SOAP Process 123
Toolkit Application 86 Deciding Which Application Modules
Problems You’ll Experience 87 to Change 126
Using the Microsoft SOAP Avoiding Protocol-Related Problems in
Toolkit 90 Modified Applications 129
Integrating New Modules with Existing
An Overview of the Application 98 Application Elements 134
How SOAP Applications Differ 98 Implementing SOAP with COM
Basic Application Design and Data Language Binding 136
Flow 99 Productivity Tips 137
Understanding the Simple Application
in This Chapter 100 Updating a Simple Utility Program 140
Updating the Server-Side
Shortcuts for Creating SOAP Component 141
Applications Quickly 100
Creating the Local Component 142
Understanding Namespaces, the Short Updating the Standard Client 144
Version 102
Updating a Data Viewer 145
Creating the Server Side Code 104 Understanding the Importance of
Designing the Component 104 Three-Tier Programming 146
Designing a Listener 105 Updating the Common Business
Logic Component 147
Creating the Client Code 106 Separating the Data Viewing Logic
from the Main Database
Testing the SOAP Application 109
Component 148
Checking the Example Application
Creating the Data Viewer Client 151
for Errors 110
SOAP Testing Tips 112 Updating a Complete Database
Testing Your Server 113 Application 153

Handling SOAP Errors 116 Modified Application Concerns 155

Reliability 156
Performance Concerns for all
Security 157
Applications 117
Performance 158
Attributes Versus Elements 118
Code Optimization 119 Troubleshooting 159

Project 120
 Contents v

6 Creating Remote Access Utilities 161 Empty and NULL Value

Processing 202
Introduction 162 Unregistering Your Control 203
Understanding CDATA Sections 204
An Overview of Remote Access
Utilities 162 Understanding XML Document
Uses for Remote Access Utility Transmission Restrictions 204
Applications 163 Using a Third-Party Product to
Understanding the Web Services Document Your Components 206
Difference 165 Creating a Simple Survey Form 208
When Is SOAP Overkill? 166 Creating the Database 210
Making Utility Programs Flexible 167 Creating the Server-Side
Shortcuts for Using Existing Component 210
Components with Utilities 168 Designing a Survey Form 212
Utility Program Security Issues 169 Testing the Input Application 214
Non-Issues for Utility Programs 170 Writing an Analysis Component 216
Writing a Server Status Viewer 172 Designing an Output Application 218
Creating the Server-Side Survey Application Data Processing
Component 173 Concerns 221
Tips for Working with Server Status Testing the Output 221
Information 177 Creating a Simple Data Entry Form
Working with the ISAPI Listener 179 Application 221
Creating the Client 181 Creating the Database 223
Designing the Server-Side
Creating a Simple Employee Check-In
Application 182 Component 223
Creating the Component 183 Creating a Data Entry Client 225
Some Caveats About WSDL Testing the Data Entry
Files 188 Application 227
Creating the Client 190 Security, Privacy, Performance, and
Testing the Application 191 Reliability Issues 227
Security 228
Project 193
Privacy 229
Creating Your Own Utility 193
Performance 230
Upgrading the Server State and User
Reliability 231
Check-In Utilities 194
Handling Data Entry and Survey Form
7 Creating Data Entry Forms Errors 232
and Surveys 195
Using Templates for Quick Forms 233
Determining Which Data Entry Vehicle
to Use 197 Using MIME for SOAP
Applications 234
Shortcuts for Data Entry and Survey Special Considerations for MIME
Applications 199 Messages 234
Reducing the Number of Round Where MIME Support Is Going 235
Trips 199
Choosing Document or RPC Style Project 236
WSDL Files 200
vi Special Edition Using SOAP

8 Providing Remote Database Access 237 Troubleshooting 279

How Do I Detect SOAP-Related
Remote Database Application Uses and Database Errors? 279
Concerns 239
Are the Database Shortcomings of
Concerns 240
SOAP Permanent? 280
Common Uses Based on SOAP
What Are the Top Ten Issues for SOAP
Limitations 242
Database Developers? 280
Developer Shortcuts for Remote
Database Applications 243 9 Moving to Web-Based Applications 283

Using Complex Data Types 245 Uses for Web-Based Applications 285
Understanding the Problem 245
Overcoming Problems with Web-Based
Understanding WSDL Generator
Applications 287
Differences 247
Describing the Interface 249 Updating a Thick Client Application for
Creating the Server-Side Thin Client Use 289
Component 251 Generating Proxy Classes the Easy
Creating the Remote Client 252 Way 290
Testing the Complete Data Type Using Thick and Thin Clients
Application 254 Simultaneously 291
Creating the Server-Side
Defining the SQL Server Database 255 Component 292
Creating the Server-Side Creating the Processing
Component 255 Component 295
Tips for Working with Database Designing the Thick Client
Components 256 Application 297
Working with SQLXML 257 Designing a Thin Client Form
Using Multiple Server-Side View 298
Components 258 Designing the Thin Client Web
Generating the Code 259 Page 299

Creating a Middle-Tier Component 263 Creating a Live Data Application 302

Creating the Client-Side Application 267 Handling Web-Based Application

Errors 303
Testing the Application 272 Handling Connection Loss 304
Scripting Errors 304
Quick Fixes for Remote Database Component Security Problems 305
Applications 274
Service Is in Use 305
Loss of Connection 275
ActiveX Component Can’t Create
Odd Data Entry Errors 275
Object 306
Performance Issues 276
Nothing Happens or Strange Error
Server Is Busy or Missing Objects 277
Message 307
Addressing Transaction Issues 278
 Contents vii

Security Issues for Web-Based Server Differences Revisited 337

Applications 307 Testing the Application 338
An Overview of the Potential Security
Solutions 309 Addressing PDA Display Issues 338
Using SSL 311 Screen Size 339
Using IBM Web Services Toolkit 311 Using Color 340
Pointer Pointers 340
Quick Fixes for Memory and Other
Resource Problems 312 Beyond PDAs to Telephones 341
Component Interactions 313
Understanding PDA Security Issues 341
Benefits of the Script Debugger 314
Human Language Support 316 Troubleshooting 343
ASP and Component Why Does the Example Seem to
Communication 316 Run, and Then Display Nothing
Component Doesn’t Support Locale Onscreen? 343
Error 317 Why Doesn’t My Code Run Properly
on All of My PDAs? 344
Case Study 317
How Do I Fix Messaging Problems
An Overview of the Problem 317
with the Client? 344
The Solution 318
The SOAP Connection 319
A SOAP Data Types and Data Type
A Negative to Consider 319
Conversions 345
10 Working with PDAs 321 Data Types Overview 346
Special Needs for PDAs 323 Complex Data Types 349
The Case for PDAs 323
Special Add-ons 324 Differences in Implementation 351
Networking 325
Data Type Conversions 353
Operating System 325

Getting SOAP for Your PDA 326 B Microsoft BizTalk and SOAP 355
pocketSOAP 327
What Is BizTalk? 357
IdooXoap 328
kSOAP 328 How BizTalk and SOAP Work
Together 358
Updating the Complex Type
Example 329 An Overview of Useful BizTalk Utilities
Creating the Client Code 329 for SOAP 359
Differences in Implementation 331 BizTalk Editor 359
Testing the Application 331 BizTalk Mapper 363
BizTalk Orchestration Designer 365
Updating the Computer Name
Example 332 BizTalk Fixes a Few SOAP Problems 368
Creating the Code 333
A Look at the Message Traffic 335 Is BizTalk the SOAP Add-On for Your
Company? 369
viii Special Edition Using SOAP

C Third-Party Tool Reference 371 An Overview of the Application 394

Finding the Right Tools 372 Creating the Server-Side

Tools That Make You Productive 372 Component 395
Tools That Fix Problems 373 Generating a Static WSDL File 395
Developing the Server-Side
Masker 2.0 374 Component Code 396
Making the CONFIG.XML
MZTools Add-In for Visual Basic 376
Entry 399
psWSDL Wizard 381
Creating the Client 400
tcpTrace 383 Initial Setup 400
Adding the Code 401
XML Spy 384
Typical File Views 385 Handling SOAP Errors 404
Viewing Data in More Than
One Way 386 Glossary 407
Special Features 388
Index 427
D SOAP for Visual C++ Developers 389

Introduction 390

An Overview of the 4S4C SOAP

Toolkit 391
Features 391
Installation 393
About the Author
John Mueller is a freelance author and technical editor. He has writing in his blood, hav-
ing produced more than 50 books and more than 200 articles to date. The topics range
from networking to artificial intelligence and from database management to heads-down
programming. Some of his current books include several COM+ developer guides, a small
business and home office networking guide, and a Windows 2000 Performance, Tuning,
and Optimization book. His technical writing skills have helped more than 25 authors
refine the content of their manuscripts. John has provided technical editing services to both
Data Based Advisor and Coast Compute magazines. He has also contributed articles to maga-
zines like SQL Server Professional, Visual C++ Developer, and Visual Basic Developer.
When John isn’t working at the computer, you can find him in his workshop. He’s an avid
woodworker and candle maker. On any given afternoon, you can find him working at a lathe
or putting the finishing touches on a bookcase. One of his newest craft projects is glycerin
soap making, which comes in pretty handy for gift baskets. You can reach John on the
Internet at [email protected]. John is also setting up a Web site at http://www.mwt.net/
~jmueller/. Feel free to take a look and make suggestions on how he can improve it. One of
his current projects is creating book FAQ sheets that should help you find the book informa-
tion you need much faster.
Dedication
This is my 50th book, so I spent time considering everything that has gone on during the past
13 years. So many people have helped me that listing them here would consume many pages.
Authors require help from those around them. They not only require technical help, but also
help with other issues such as the daily needs that we all have. I dedicate this book to all of those
tireless helpers who have meant so much to me over the years. I wish that I could list all of you
on this page.

Acknowledgments
Thanks to my wife, Rebecca, for working with me to get this book completed. I really don't
know what I would have done without her help in proofreading my rough draft. She also
helped research, compile, and edit some of the information that appears in this book.
Russ Mullen deserves thanks for his technical edit of this book. Russ greatly added to the
accuracy and depth of the material you see here. In addition, he spent many hours working
with me through e-mail to find solutions to some of the problems that SOAP presented,
and he's responsible for providing many of the Web-site addresses sprinkled throughout the
book. Russ went even further in this book; he spent hours finding just the right equipment
to use when performing his technical edit. I can never sufficiently express my thanks to him.
As you read this book, you'll see mentions of many third-party products and in-depth looks
at products from major vendors. The SOAP community provided much of this information
through conversation and by checking the technical accuracy of both code and theory. I'd
like to thank everyone, but space won't allow me to mention them all. However, I would
like to mention that the people on the Developmentor list server and the Microsoft SOAP
newsgroups were exceptionally helpful.
Some special people stand above the rest. Simon Fell helped me in many areas, including
the complex type and PDA examples. He also helped with the products he develops and
supports, including 4S4C and pocketSOAP. Roger Wolter helped in many of the Microsoft
SOAP Toolkit examples. He read and commented on many of the theory sections of the
book. Paul Kulchenko discussed MIME and SOAP attachments with me, among other top-
ics. Karen Watterson provided several tips for the book and pointed me toward some of the
third-party products. Rosimildo DaSilva taught me about SOAP and embedded systems.
Yasser Shohoud helped me understand the problems with user-defined type support in
SOAP toolkits. Birgit Reidinger helped me with XML Spy and discussed some compatibil-
ity issues with me. Ahmad Baitalmal provided me with a glimpse of a completely functional
SOAP application that appears as one of the case studies in this book. He reviewed and
helped me refine the case study so that you could read about a real-world application in
action. Jacek Kopecky provided me with information about IdooXoap and discussed PDA
issues with me.
Every book begins with a good outline and proposal. I would like to thank the following
reviewers for their help in getting this book off to a good start: Clemens Vasters, Jeremiah
Talkar, Ken Rabold, Madhu Govinda, Rob Caron, and Alek Aslom. All of these reviewers
provided solid comments that affected the final content of the book and my approach in
discussing certain topics.
Matt Wagner, my agent, deserves credit for helping me get the contract in the first place
and taking care of all the details that most authors don't really think about. I always appreci-
ate his help. It's good to know that someone wants to help.
Finally, I would like to thank Candace Hall, Michelle Newcomb, Sarah Robbins, Karen
Shields, Maureen McDaniel, and other members of the Que staff for their assistance in
bringing this book to print. Writing a book about SOAP presented many logistical chal-
lenges, and I appreciate their willingness to give me the time required to put a good book
together. Que was also kind enough to provide me with a Palm to use in the PDA section of
this chapter.
Tell Us What You Think!
As the reader of this book, you are our most important critic and commentator. We value
your opinion and want to know what we're doing right, what we could do better, what areas
you'd like to see us publish in, and any other words of wisdom you're willing to pass our way.
As an associate publisher for Que, I welcome your comments. You can fax, email, or write
me directly to let me know what you did or didn't like about this book—as well as what we
can do to make our books stronger.
Please note that I cannot help you with technical problems related to the topic of this book, and that
due to the high volume of mail I receive, I might not be able to reply to every message.
When you write, please be sure to include this book’s title and author as well as your name
and phone or fax number. I will carefully review your comments and share them with the
author and editors who worked on the book.
Fax: 317-581-4666
Email: [email protected]

Mail: Associate Publisher

Que
201 West 103rd Street
Indianapolis, IN 46290 USA
 INTRODUCTION

In this chapter
What's in This Book 3
On the Web 6
Intended Audience 7
Equipment Used for This Book 7
Conventions Used in This Book 8
2 Introduction

It is difficult to get some types of applications to work on the Internet because of the lack of
technologies that can move data from one point to another safely and with few technical
problems. In addition, developers need a single standard for transferring data that will work
on all operating system platforms. The need for a single solution for data transfer is becom-
ing apparent as financial institutions begin moving their operations to the Internet.
Recently, credit card companies began actively promoting their Web site as a better alterna-
tive for customer support than calling on the telephone. Many trade journals have also had
stories of major financial institutions making the move to the Internet.
DCOM and CORBA, the two most popular technologies currently in use, have a myriad of
technical problems, including lack of operating system platform independence. The fact
that many Web servers with firewalls won’t work with either technology only complicates
matters. In addition, the use of two technologies means that developers end up working
twice as hard to get their distributed applications to work.
Clearly, developers require a new data transfer technology, and I believe that technology
will be the Simple Object Access Protocol (SOAP). SOAP is based on XML and allows data
transfer through firewalls with a minimum of problems. Because SOAP is XML-based,
developers find that it suffers few platform-related problems. (Of course, compatibility is
always an issue with a new technology.) In short, SOAP is the technology that you need to
solve your distributed application programming problems. It will allow you to transfer data
between disparate machines more efficiently and with fewer problems.
SOAP is such a new technology that most developers have to rely on the white papers pro-
duced by developers of the technology for reference purposes. Special Edition Using SOAP
provides the kind of information that you’ll need to get your applications working quickly
after you understand the basic theory behind SOAP. This book will help you gain
proficiency in SOAP in the following ways:
■ A technology overview that cuts out superfluous information and provides just the
information needed to write applications
■ Detailed examples that show how to create new applications as well as convert
existing ones
■ Tips that show how to become proficient with SOAP quickly
■ Solutions for problems with using SOAP in certain environments
■ Performance and other tips that will make applications perform faster and more
reliably
■ A data type and data conversion reference
■ An overview of Microsoft’s Biztalk technology
■ A third-party tool reference that will allow you to start working with SOAP immediately,
rather than build tools first.
 What’s in This Book 3

As you can see, this book provides complete coverage for the developer who needs to com-
plete projects quickly. Special Edition Using SOAP focuses on getting work done, rather than
on the theoretical aspects of the technology. Of course, the book covers all of the details,
such as ensuring that security is in place and that the application will perform reliably.
Although the focus is on speed, the book will also provide tips that ensure development
speed gains aren’t overshadowed by hours of debugging later. Special Edition Using SOAP is
the book you need on your shelf because it provides the real-world tools that you require to
make Web-based applications a reality.

What’s in This Book

I’ve designed this book to get you up and running with SOAP as quickly as possible. In fact,
this book contains only three theory chapters—the remaining seven chapters focus on
example code and how to accomplish work quickly. You’ll find a wealth of productivity aids
and tips on how to avoid pitfalls. The following sections provide an overview of the book.

Chapter 1: An Overview of SOAP

This chapter tells you what SOAP is and how it differs from older technologies, such as
DCOM and CORBA. It also talks about how SOAP uses HTTP and XML with some addi-
tional elements to create a message. Finally, we’ll talk about the advantages and disadvan-
tages of using SOAP to develop applications. This includes problems you might experience,
such as getting the same performance from your application as before you converted it to
use SOAP.

Chapter 2: SOAP in Theory

SOAP is a protocol (set of rules), rather than an application programming interface (API) or
programming methodology. This chapter will help you understand how SOAP works and
why it can avoid certain problems that DCOM and CORBA can’t in a distributed environ-
ment. The coverage in this section will be complete, but extremely focused. The chapter
provides many references to Microsoft and other vendor material on the Internet. Even if
you need a detailed view of the inner workings of SOAP, the combination theory overview
in this chapter and references to white papers on the Internet should provide everything
you need.

Chapter 3: An Overview of Security Issues for SOAP

Security issues are a concern for any application. Protecting data is a problem that has
plagued computers since the very beginning. The distributed application environment
provided by the Internet only makes matters worse.
4 Introduction

This chapter discusses the major security concerns for any SOAP application. Later chap-
ters will discuss specific concerns for particular application types. You’ll also find a discus-
sion of privacy issues in this chapter. Securing the user’s identity is becoming more
important as countries pass laws requiring secure and hassle-free communication.
Finally, this chapter discusses some alternative technologies, such as biometrics, for securing
your system. It’s important to look at these alternatives as a way to reduce coding require-
ments and application usage complexity.

Chapter 4: Using SOAP to Create a Simple Application

This chapter contains the first complete SOAP example in the book. The example isn’t that
complicated, but we do examine it thoroughly. The chapter will discuss issues such as error
handling and testing techniques. We’ll also talk about methods for creating WSDL files and
discuss how those WSDL files reflect the underlying component technology. This chapter
concentrates on the Microsoft SOAP Toolkit.

Chapter 5: Migrating an Application from DCOM to SOAP

Many developers will spend more time converting existing applications to SOAP than cre-
ating new ones. This chapter shows how to move an application from the DCOM environ-
ment to the SOAP environment. It also shows how to maximize your current coding
investment by continuing to use as many of the DCOM modules as possible. We’ll also
discuss some practical issues. For example, SOAP isn’t always the correct solution to your
application problem; you might find that DCOM is still the right tool for the job.

Chapter 6: Creating Remote Access Utilities

More employees are on the road than ever before. They want access to their applications
back at the office and it’s your responsibility to provide that access. In addition, a remote
user might have new application needs, such as a utility for quickly checking in at the home
office without calling. This chapter looks at these issues and more. You’ll find two program-
ming examples that show how to use SOAP for the remote user’s needs.

Chapter 7: Creating Data Entry Forms and Surveys

Web-based applications are useful because you can share information with others outside
your company. You can also gather information from others and use it for analysis. For
example, companies are always looking for new ways to request information from customers
and use it to improve sales or service.
SOAP is a useful technology because it extends the data entry form and survey application
to perform tasks that you can’t generally perform now. For example, such an application can
completely automate the process of entering new data into the database. For that matter,
you can automate the analysis as well and send the results to concerned parties.
 What’s in This Book 5

This chapter also discusses some practical issues you need to consider, such as using tem-
plates to reduce the amount of coding required to create an application. Templates are espe-
cially important for survey applications where the presentation changes regularly. We’ll also
discuss some special issues, such as the enhanced error-handling requirements for a survey
application.

Chapter 8: Providing Remote Database Access

The database is the center of every business. In fact, the organized storage of data is the
center of every business, even those that don’t have a computer (admittedly rare today).
SOAP provides new opportunities to handle data in the distributed application environ-
ment. Not only can you create new data connections internally, but also partners can now
access your data directly as needed.
This chapter addresses some special database applications needs. For one thing, most devel-
opers want to use transactions to ensure reliable data transfer from one point to another.
We’ll discuss this issue along with methods for fixing database application problems.
Although this chapter won’t provide you with a complete tutorial on database design and
implementation, it does provide you with the SOAP piece of the puzzle.

Chapter 9: Moving to Web-Based Applications

Several of the previous chapters have led up to this one. The application of the future will
always provide a distributed connection. The user won’t care where he uses the application
because it won’t matter. Web-based applications represent the future of programming tech-
nology. This chapter shows you how to create some simple Web-based applications.
Of course, the important issue is preserving all of that code you have right now. This chap-
ter also tells you how to preserve your investment by leveraging your current code base.
You’ll avoid creating applications from scratch by carefully designing your Web-based
application to hook into the existing application infrastructure.

Chapter 10: Working with PDAs

Many users are completely hooked on the personal digital assistant (PDA). This little device
holds a significant piece of their lives. It stores all of their contact information, tasks they
need to perform, and schedule for upcoming weeks. Unlike a desktop computer, a PDA is
small and you can take it with you anywhere to record new information. In short, PDAs are
becoming an essential asset for most employees.
It won’t be long before you’ll find yourself moving existing applications to the PDA—at
least the applications that are small enough to move. This chapter examines methods for
moving common applications to the PDA. We’ll look at two examples so that you can see
some of the pitfalls in developing a PDA solution.
6 Introduction

This chapter also discusses some of the issues surrounding PDA development. We’ll cover
SOAP-specific issues, such as the selection of a SOAP toolkit for PDA development. This
chapter also discusses a few general PDA issues. For example, the small screen that a PDA
provides will affect the way that you develop the application interface.

Appendix A: SOAP Data Types and Data Type Conversions

This short appendix tells what data types SOAP supports and how to convert data from
existing languages to a format that SOAP will understand. It’s a quick reference with code
snippets that show how to perform the required work. There aren’t any complete program-
ming examples.

Appendix B: Microsoft Biztalk and SOAP

Although Microsoft isn’t the only company working with SOAP, for readers of this book,
Microsoft will probably be a major supplier of SOAP technology. Biztalk provides access to
Microsoft’s XML and SOAP technology implementations, so a discussion of this important
server product is essential. This appendix provides an overview of Biztalk as a whole, plus
specifics on how this server will provide XML and SOAP services.

Appendix C: Third-Party Tool Reference

Because SOAP is so new, you’ll want to know what types of tools are available from third
parties. This appendix provides a list of the tools that are available at the time of writing.
This list isn’t complete, but it does represent some of the better tools available for SOAP
development today.

Appendix D: SOAP for Visual C++ Developers

Most of the examples in this book concentrate on Visual Basic or some form of scripting.
This appendix discusses SOAP for the Visual C++ developer. You’ll find example code and a
complete discussion of Visual C++ development issues. Like the example in Chapter 4,
“Using SOAP to Create a Simple Application,” the appendix example is somewhat simple,
but the discussion is in depth. We’ll talk about issues such as testing and SOAP error han-
dling, as well as the creation of client-side code.

On the Web
We’ve provided all the source code for the examples in the book at an easy to find website.
Just go to www.quepublishing.com and check out the site. There you will find easy to down-
load executables that contain all the source code separated by chapter. Everything you need
to complete the exercises is right there.
 Equipment Used for This Book 7

Intended Audience
I wrote this book for the developer who has a good understanding of programming princi-
ples and has worked with distributed applications sometime in the past. In other words, a
complete novice will have a difficult time understanding the material, but someone with a
little experience will get some information from it almost immediately. All of the examples
assume that you know something about Visual Basic or Visual C++ programming and have
worked with DCOM or CORBA in the past. You need to understand terms like interface
because the theory sections in this book are short and concentrate only on SOAP principles.
As the book progresses, the topics become more difficult and the anticipated understanding
level of the reader increases. By the time you reach the remote database programming
example (Chapter 8, “Providing Remote Database Access”), I’m assuming that you’re at
least an intermediate to advanced reader who has spent some time developing remote data-
base applications. You must understand how to work with databases. Although the chapters
do provide details on constructing the applications, you won’t find any information about
working with the database managers or procedural steps for constructing the tables. All that
I provide is the schema you need to build the application.
This doesn’t mean that I’m going to bury you with arcane information; this book covers the
various topics in simple terms that you’ll find easy to understand. The purpose of this book
is to discuss as many SOAP issues as possible in terms that everyone will understand.
However, a beginner will still very likely get lost as the book progresses because there’s little
in the way of introductory information.

Equipment Used for This Book

I made some assumptions while writing the application programming examples in this book.
First, you need at least two machines: a workstation and a server. This two-machine setup is
the only way that you’ll see SOAP in action and truly know it works as anticipated. During
the writing of this book, I used a Windows 2000 and Windows 98 workstation. The servers
included Windows 2000 Server and Red Hat Linux. I also experimented with two Web
servers running on a NetWare file server, and two PDAs (Windows CE-based Pocket PC
and a Palm).
I tested all of the examples in this book with Visual Studio 6.0 Enterprise Edition. Most of
the examples use Visual Basic, but you’ll also find examples written in Visual C++ and in
various scripting languages. None of these examples is guaranteed to work with any other
programming language products and none of them will work with the educational versions
of Visual Studio.
8 Introduction

You must install the latest service packs for all products before the examples will work prop-
erly. SOAP is a new technology that relies on the latest versions of many DLLs, especially
the XML parsers used in the examples. This book doesn’t support the beta versions of the
SOAP toolkits unless I specifically note that I’ve used a beta version. Make sure you use
release versions of the SOAP toolkits, especially the Microsoft SOAP Toolkit Version 2.0.
Some of the example programs rely on a database manager. I used SQL Server 7.0 for all of
the examples in this book. The source code provides scripts that I tested on SQL Server,
but may work with other database managers as well. The sample data is in delimited text
format and you should be able to import it into any database manager. In short, you can use
any database manager you want, but the examples might require modification to do so.

Conventions Used in This Book

There are several conventions used within this book will help you get more out of it. Look
for special fonts or text styles and icons that emphasize special information.
■ Sometimes I’ll ask you to type something. This information always appears in bold type
like this: Type Hello World.
■ Code normally appears on separate lines from the rest of the text. However, there are
special situations where small amounts of code appear directly in the paragraph for expla-
nation purposes. This code will appear in a special font like this: Some Special Code.
■ Definitions are always handy to have. I’ll use italic text to differentiate definitions from
the rest of the text, like this: A CPU is a required part of your machine.
■ In some cases, I won’t have an exact value to provide, so I’ll give you an idea of what
you should type by enclosing it in angle brackets like this: Provide a <Machine Name>
value for the Name field.
■ You’ll always be able to recognize menu selections and command sequences because
they’re implemented like this: Use the File | Open command.
■ URLs for Web sites are presented like this: http://www.microsoft.com.

Notes help you understand principle or provide amplifying information. In many cases,
a note is used to emphasize some piece of critical information that you need.

All of us like to know special bits of information that will make our job easier, more
fun, or faster to perform. Tips help you get the job done faster and more safely. In
many cases, the information found in a Tip is drawn from experience, rather than
through experimentation or the documentation.
 Conventions Used in This Book 9

Any time you see a caution, make sure that you take special care to read it. This infor-
mation is vital. I’ll always uses the Caution to designate information that will help you
avoid damage to your application, data, machine, or self. Never skip the Cautions in a
chapter, and always follow their advice.

Finding what you need quickly is more important than ever before. Many people work
at a pace that they call Internet time. They no longer have the luxury of performing
hours of research to find that magic piece of information needed to complete an appli-
cation. With this in mind, I’ve spent the time you don’t have to find unique information
sources on the Internet. This icon will always provide that information so that you can
get what you need quickly.
 CHAPTER
1
An Overview of SOAP
In this chapter
What Is SOAP? 13
How SOAP Differs from DCOM and CORBA 15
SOAP, HTTP, and XML 22
Problems Solved by Using SOAP 24
Performance Issues 26
SOAP and the Web Server 28
Why Make the Move to SOAP? 28
Case Study 29
12 Chapter 1 An Overview of Soap

At one time in the history of the PC, computing consisted of a single computer using simple
applications that relied only on local resources. Eventually, networking changed the way
people shared expensive peripherals and data, but the data was still, to an extent, stored
locally. As companies grew and became more dependent on the PC, the concept of the
Local Area Network (LAN) gave way to the Metro Area Network (MAN) and Wide Area
Network (WAN). The client/server architecture of days gone by allowed a limited form of
distributed computing. However, there was still a direct connection and all of the data
appeared within the confines of one company, so sharing data was still relatively easy.
Today, computing is all about distributed applications running on machines that may not
know anything about each other. Data is no longer restricted to one company; business-to-
business communication is now the norm. Consequently, these machines may not even have
a direct connection or access the network at the same time. Data sharing occurs between
individuals who may not ever meet. In short, there isn’t a direct connection between the
provider and the user of data anymore, which means the rules used in the past don’t work
well for modern communication needs.
Early computers relied on protocols (predefined rules) that ensured safe data transfer
between machines that knew what to expect from each other. These protocols work fine on
a LAN, MAN, or WAN because there’s a direct connection between machines. A protocol
designed for the LAN environment, however, may run into problems when dealing with
something like the Internet. For example, the protocol may expect to find another Windows
machine when the client really needs to communicate with a Unix server. That’s precisely
what’s happened and why we need a new protocol named Simple Object Access Protocol
(SOAP). The fact that you’re reading this book means that you already have some idea of
why you need SOAP and may even know something about it from a technical perspective.
The first section of this chapter is going to tell you more about SOAP—what it does and
how it works. This section isn’t going to provide many details, but it will prepare you for the
detailed discussion in Chapter 2. I want to present a basic overview of the technology before
jumping into the details.
SOAP is different from older protocols. It provides some features these older protocols
don’t have. For example, unlike Distributed Component Object Model (DCOM) and other
binary protocols, SOAP won’t interfere with the operation of the firewall on your Web
server. It uses a plain text transfer method that firewalls readily accept. This means that you
can maintain the integrity of your company’s Web site, and still get the data you need.
Some of these neat new features come at a cost. You lose some of the features provided by
the older protocols. For example, security is an issue that many SOAP developers are trying
to address as of this writing. The second section of the chapter provides a brief comparison of
SOAP to older protocols, such as DCOM and Common Object Request Broker Architecture
(CORBA). You’ll learn how SOAP excels and where it provides less than stellar results.
SOAP actually interacts with other Web technologies that you may have used in the past.
Although SOAP can theoretically rely on any transport protocol, the current toolkit from
Microsoft uses the same Hypertext Transfer Protocol (HTTP) used to move Web pages
across the Internet. (Note that you can use any reliable protocol to transfer SOAP
 What Is SOAP? 13

messages—HTTP is simply the most straightforward and easily accessible method right
now, so developers are using it.) In addition, SOAP relies on the eXtensible Markup
Language (XML) to format data before moving it from one point to another. The third
section of the chapter discusses the interaction between these three technologies.
Ch
As previously mentioned, Microsoft designed SOAP to address specific problems. Technologies
such as DCOM just can’t make the grade in the distributed computing environment of the 1
Internet. The fourth section of the chapter discusses these problems in detail and explains how
SOAP addresses them.
SOAP changes the performance picture—it uses a new technique to transfer data, so differ-
ences in performance are expected. In many cases, SOAP is slower than older technologies
like DCOM. You can’t beat a direct binary connection between client and server that relies
on optimized data transfers. However, given the distributed nature of applications today,
there are also situations when SOAP is faster than binary technologies. In the fifth section
of the chapter, we’ll talk about SOAP performance issues. This section also talks about ways
of mitigating some of those performance losses by using better coding techniques.
Microsoft designed SOAP to work with the Internet—which means you’ll eventually run
into a Web server. The sixth section of the chapter discusses how SOAP works with a Web
server to handle client requests. We’ll talk about some of the mechanics you’ll need to know
later, like common storage locations for files and some of the implementation details.
The final section of the chapter discusses a topic that many of you will ask about once you
see everything required to move to SOAP. It’s important to know why this move is so
important and how you’ll benefit from it. SOAP is a great technology with a very important
purpose for your company. This section discusses why you need to include SOAP in your
programming toolbox.

What Is SOAP?
SOAP is a lightweight communication protocol based on the eXtensible Markup Language
(XML). It allows applications and components to exchange data. As mentioned in the intro-
duction, SOAP currently relies on HTTP as a transport protocol, but could use any reliable
transport protocol to transfer data. In addition, SOAP is useful for many data transfer needs,
including LANs, WANs, and MANs—it’s not just for the Internet.

Learning XML is an essential part of learning SOAP. You don’t have to become an XML
guru, but knowing the basics is a requirement. Many XML Web sites offer tutorials and
other information about this technology. For example, DevelopMentor (http://www.
develop.com/dm/dev_resources.asp) provides good tutorials on this and other
distributed application topics from a Microsoft perspective. W3Schools (http://www.
w3schools.com/ xml/) provides detailed tutorials in small segments. Another good
place to look for XML training is Courses in XML by QTrain (http://www.qtrain.
net/). Once you finish the basic XML tutorials, you’ll want to visit XSLT.com (http://
xslt.com/resources_tutorials.htm) and learn more about data transformation
14 Chapter 1 An Overview of Soap

techniques. ZVON.org (http://www.zvon.org/index.php?nav_id=2) includes

tutorials on both XML namespaces and the eXtensible Stylesheet Language
Transformations (XSLT).

Despite what you may have heard, SOAP isn’t another conspiracy by Microsoft to take over
the world—it’s a protocol supported by many vendors. Some of the most notable contribu-
tors to SOAP are Ariba, Commerce One, Compaq, DevelopMentor, HP, IBM, IONA,
Lotus, SAP, and UserLand. Vendors who support SOAP hope it will eventually gain stan-
dards status. The World Wide Web Consortium (W3C) is currently discussing SOAP. You
can read the W3C comments at http://www.w3.org/Submission/2000/05/Comment and see
the initial specification at http://www.w3.org/TR/SOAP/.
Since SOAP is a new protocol that isn’t tied to a particular operating system, the vendors
working on it are free to add features that make SOAP especially suited to distributed
application use. Three of the features that make SOAP attractive are:
■ No ties to existing component technologies—SOAP will theoretically work with
any platform.
■ No ties to a particular programming language; you can use SOAP with any language
capable of outputting text. (You can also use a special toolkit to output the formatted
text as well, seen in Chapter 4.)
■ Easy to learn and simple to extend.

You’ll see as the book progresses that these three points are important because they’ll affect
your perception of SOAP as an application solution. For example, the fact that vendors haven’t
tied SOAP to a particular programming language means that you’ll very likely see a wealth of
toolkits on the market. These toolkits may not all produce compatible code. In addition, they
may rely on server or client-side components that aren’t compatible with other SOAP imple-
mentations. Problems like these will become less noticeable as SOAP nears standardization.
The last point is important as well. SOAP really is easy to learn. It’s verbose, which means
that some of the code listings you see become quite long and look complicated, but the
underlying technology is simple. The complexity that you’ll see as we work with SOAP
throughout the book comes from the various extensions that Microsoft and other vendors
add. These extensions provide additional flexibility and allow you to tailor a SOAP imple-
mentation to specific needs.

SOAP is a very hot topic right now because it offers so much. However, getting the latest
news can prove difficult because the specifications change so quickly. You can usually rely
on news Web sites and newsgroups to provide updated information on a regular basis.
For example, the SOAP News Web site at http://soap.weblogs.com/ provides up to
the minute information about this new standard. DevelopMentor and other organizations
provide list servers that discuss SOAP. The main public SOAP newsgroups appear on the
 How SOAP Differs from DCOM and CORBA 15

Microsoft server (news.microsoft.com). You’ll find a wide range of topics in news-

groups like microsoft.public.msdn.soaptoolbox, microsoft.public.xml.
soap, and microsoft.public.xml.soapsdk.

Ch

SOAP is intertwined with other technologies as well. We’ve already discussed its connection 1
with XML (and by extension, XSLT ). Creating a SOAP implementation is only the first
step. If you want others to use your implementation, you need to advertise it in some way.
As the book progresses, we’ll talk about the significance of Universal Description,
Discovery, and Integration (UDDI). You can find out more about this technology at
http://www.uddi.org/. The main purpose of UDDI is to allow vendors to advertise
services publicly, including those that rely on SOAP.
Another technology that figures prominently in the SOAP picture is Web Services
Description Language (WSDL). We’ll discuss this technology in detail in Chapter 2.
WSDL provides a description of objects and documents on a Web server in the form of a
schema. It also allows advanced features of some programming IDEs to display a list of
functions applicable to a particular object. Look at http://www-106.ibm.com/developerworks/
library/w-wsdl.html to find out more about WSDL.

How SOAP Differs from DCOM and CORBA

SOAP, like DCOM and CORBA, is a wire protocol. The term wire protocol indicates that appli-
cations use SOAP to move data from one place to another. Many people assume that because
COM and DCOM have similar names that they’re the same kind of protocol. COM is actually
a specification that tells how to create components—DCOM enables those components to
interact across a network. It’s an important difference to understand because Microsoft designed
SOAP to overcome some of the limitations of DCOM, not to replace it or replace COM itself.
The first two sections below provide a comparison of SOAP with DCOM and CORBA.
We’ll look at how the protocols compare from a result and method perspective. The essential
difference between SOAP and the other two protocols is that SOAP uses plain text, while
DCOM and CORBA use a binary format.
A third section discusses the Java Remote Method Invocation (RMI). This is a relatively
simple wire protocol that’s designed to support Java. It differs from DCOM and CORBA
because Java RMI is only designed to support Java—these other protocols support any
language. The Java RMI is still a binary protocol, so it differs from SOAP in that regard.

DCOM Wire Protocol

The DCOM Wire Protocol, also known as the DCOM Network Protocol or DCOM for
short, is a high-level protocol based on the Distributed Computing Environment (DCE)
Remote Procedure Call (RPC) Network Protocol and implemented by Microsoft. DCOM
relies on several lower-level protocols to accomplish its work. Like SOAP, DCOM oversees
the data transfer; it doesn’t perform the mechanics of moving the data. As an enabling
16 Chapter 1 An Overview of Soap

protocol, DCOM ensures that the client and server can talk to each other, but doesn’t
participate in the conversation or manipulate the data.

Microsoft didn’t create the DCE RPC Network Protocol specification. The Open Software
Foundation (OSF), which is now part of the Open Group, created it. You can find out
more about the DCE RPC Network Protocol specification at http://www.osf.org/.
Find a good RPC overview at http://www.ja.net/documents/ NetworkNews/
Issue44/RPC.html. There’s an article about the inner workings of DCOM at http://
www.microsoft.com/msj/0398/dcom.htm. The main Microsoft DCOM Web site is
at http://www.microsoft.com/com/tech/DCOM.asp.

Figure 1.1 shows a generic DCOM component setup. We aren’t doing anything fancy here
since the idea is to learn how the connection between the client and server works. Table 1.1
tells you about the various components shown in Figure 1.1. Notice that the message goes
through quite a few transition layers. These layers figure prominently in the performance
discussion later in the chapter.

Figure 1.1 Server-

Internet Explorer Client Side
versions 5.5 and Component
above can display OLE32
SOAP messages in a
simple format.
Service Control Service Control
Proxy Stub
Manager (SCM) Manager (SCM)

COM COM
Runtime Runtime

Security Provider Security Provider

DCOM Network Protocol DCOM Network Protocol

(DCE RPC Network Protocol) (DCE RPC Network Protocol)

Protocol Stack Protocol Stack

Winsock Driver Winsock Driver

User Datagram Protocol (UDP) User Datagram Protocol (UDP)

Internet Protocol (IP) Internet Protocol (IP)

Ethernet Driver Ethernet Driver

 How SOAP Differs from DCOM and CORBA 17

Table 1.1 Components of a Typical DCOM Data Transfer

Component Description
Client Originates requests to the server for resources and support.
DCOM assumes this is a standard desktop application Ch
operating on a LAN.
OLE32 DLL containing the methods used to create an instance
1
of an object (along with a wealth of other functionality).
Service Control Manager (SCM) Creates the initial connection between the client and
server. DCOM only uses the SCM during object
creation. After the client and server establish a connection,
the SCM steps out of the picture. The SCM represents
one of the handshaking elements that we’ll talk about
later in the performance section of the chapter.
Proxy The server’s presence within the client’s address space.
The proxy is actually a table of interfaces. Windows
creates and manages it at the request of the COM run
time. The proxy allows the client to think that the server
is local, even though the server is located on
another machine.
COM Runtime Operating system elements that host objects and provide
client/server communication. The COM runtime is part
of any COM related scenario—both in-process and out-
of-process—local and remote.
Security Provider The security provider logs the client machine into the
server machine. The operating system determines the
type of security providers available. Some security
providers will also protect all data transferred between
the client and server in some way—usually using
encryption.
DCOM Network Protocol Defines a protocol for creating a connection with a
(DCE RPC Network Protocol) remote server. In addition to implementing a component
protocol, this block contains all the elements to implement
the Object Remote Procedure Call (ORPC) specification
at an application level. This is the wire protocol that’s
the target of this section of the chapter.
Protocol Stack Actual network communication requires more than just
one protocol—there are network-related protocols to
consider as well. The protocol stack consists of all the
protocols required to create a connection between the
client and server, including network specific protocols
like transmission control protocol/Internet protocol
(TCP/IP). Figure 1.1 shows a typical protocol stack
consisting of a Winsock (Windows sockets) driver, user
datagram protocol (UDP), IP, and an Ethernet driver.
Not shown is the Ethernet network interface card (NIC)
actually used to create the physical connection between
the client and server.
18 Chapter 1 An Overview of Soap

Table 1.1 Continued

Component Description
Stub The client’s presence within the server’s address space.
Windows creates and manages the stub at the request of
the COM runtime. As far as the server is concerned, it’s
working with a local client.
Server The COM object that the client has requested services
and resources from.

Figure 1.1 is typical of a binary wire protocol like DCOM and even CORBA. As you can
see, the design is somewhat complex, but works well in a LAN, WAN, or MAN environment.
The point is that DCOM is an enabling technology, not a data manipulation technology.
You can view it as a sort of traffic cop and trail guide rolled into one.
DCOM uses a specific procedure to ensure reliable communications between machines.
This procedure relies on a lot of handshaking to ensure that each phase of the data move-
ment process occurs as anticipated. Contrast this with SOAP, which sends the data and sim-
ply assumes that it will arrive at the remote location. Here are the steps that DCOM follows
to ensure a safe data transfer between machines:
1. Client issues an object creation call. The call must include both a class ID (CLSID) and
a server name (along with any information required to log onto the server). As an alter-
native, the client can issue a standard call that OLE32.DLL will resolve to a remote
location based on a registry entry, or the client can use monikers.
2. OLE32.DLL calls upon the client side SCM to create a connection to the server
machine because it can’t service the call locally.
3. DCOM creates the required packets to send information from the client to the server.
4. The server-side SCM creates an instance of the desired server-side component and
returns a pointer of the object instance to the client.
5. The server-side SCM calls upon the COM runtime to create a stub for the component
to interact with.
6. The client-side SCM calls upon the COM runtime to create a proxy for the client to
interact with.
7. The SCM returns a pointer to the proxy to the client.
8. Normal client and server-side component communications begin.

As you can see, DCOM provides a robust data communication environment that relies on a
client and server that exist at the same time and have a direct connection. It’s important to
understand that DCOM is still an optimal technology when reliability is more important than
flexibility. DCOM provides a secure and reliable environment that SOAP can provide. On the
other hand, the complexities of DCOM make it difficult to use over the Internet. There are
simply too many requirements that DCOM has to satisfy before communication can occur.
 How SOAP Differs from DCOM and CORBA 19

CORBA IIOP
CORBA and DCOM have existed side by side for an eternity in computer time. Each tech-
nology has proponents and detractors who protect their point of view with the vigor of
religious zealots. However, many developers view CORBA as simply an alternative to Ch
DCOM and therefore feel that CORBA has the same limitations. The truth is slightly
different. CORBA is more like SOAP than DCOM when it comes to design goals, even 1
though CORBA is a binary protocol.
There are two main pieces to this architecture, just like COM and DCOM in the world of
Windows. CORBA, like COM, provides a specification for component services. There are
several CORBA implementations on the market. IBM provides the System Object Model
(SOM) and Distributed SOM (DSOM) architectures. Likewise, Netscape offers the Open
Network Environment (ONE) platform. Each CORBA implementation differs a little in
low-level details, which we won’t discuss in this chapter since they aren’t important in the
overall comparison of CORBA to SOAP.
Unlike COM, the Object Management Group (OMG) designed CORBA to run on more than
one operating system. In addition, OMG looked at Internet communication as an important
CORBA feature from the outset. CORBA tends to perform better than DCOM on the
Internet, but still provides less than acceptable performance in most cases. These design
differences make CORBA more like SOAP than DCOM when it comes to design goals.

You can find out more about OMG at http://www.omg.org/. One of the best Web
sites to visit for CORBA details is Distributed Object Computing with CORBA Mid-
dleware (http://www.cs.wustl.edu/ ~schmidt/corba.html). This site
includes a copy of the CORBA specification, a tutorial, and some great overviews of the
technology. There’s an interesting CORBA frequently asked questions (FAQ) site at
http://www.aurora-tech.com/corba-faq/. There are FAQs for all the CORBA
and IIOP components on this site, including answers on IIOP elements like the
Interoperable Object Reference (IOR). The OMG sponsored ORB Interoperability
Showcase appears at http://corbanet.dstc.edu.au/.

CORBA, like COM, is a local component implementation. It requires a wire protocol to

cross machine barriers. That’s where the Internet Inter-ORB Protocol (IIOP) comes into
play. Unlike DCOM, a protocol designed for LAN use alone, OMG developed IIOP for
Internet use. IIOP, like DCOM, is the wire protocol that enables component communica-
tion between remote applications. In fact, you could almost replace the block names in
Figure 1.1 with CORBA and IIOP equivalents and see the technology in operation—the
ideas are the same, the names and precise implementation details are changed.

If you want to know how the DCOM and CORBA technologies differ from a block dia-
gram perspective, compare the CORBA block diagram (Figure 1.2) at http://www.
cs.wustl.edu/~schmidt/corba-overview.html with Figure 1.1 in this chapter.
20 Chapter 1 An Overview of Soap

You’ll see definite similarities in the two architectures. For example, you can view the
COM proxy and stub performing a function similar to the CORBA Interface Definition
Language (IDL) stub, Dynamic Invocation Interface (DII), IDL skeleton, and Dynamic
Skeleton Interface (DSI). Note that CORBA uses more layers to accomplish the same
goals as DCOM—some developers say this makes CORBA slower than DCOM even on a
LAN (your mileage may vary).

IIOP provides most of the same functionality of DCOM. Unlike SOAP, both IIOP and
DCOM can transfer a variety of non-ASCII data types, such as integers and objects. The
same binary features that restrict IIOP and DCOM from getting through firewalls allow
them to transfer native data types. This makes binary data transfers more efficient—they
require fewer data packets.
CORBA and DCOM are both stateful protocols. A user establishes a connection with the
server and that connection remains in place during the entire session. This ensures the user’s
state information remains intact and enhances the developer’s ability to write applications
that interact with the user. HTTP is a stateless protocol. Consequently, SOAP can’t main-
tain state information, which causes problems. For example, SOAP can’t support properties
because the use of properties infers maintained session state information; something that
HTTP can’t support.
Another way that CORBA and DCOM are similar (although definitely not equal) is in the
area of security. If you want to secure a SOAP data transfer, you need to rely on a third party
product,such as Secure Sockets Layer (SSL). CORBA and DCOM both include built-in
security. Unfortunately, security settings often cause problems for both CORBA and
DCOM developers. Getting security set high enough so that no one can see your data, yet
low enough to get past firewalls is difficult. In addition, DCOM is especially susceptible to
security-setting errors that result in a loss of communication. If the server and the client
don’t agree about the level of security to support, then communication doesn’t take place.
Obviously, CORBA and DCOM are quite different internally. The packets they produce are
not compatible, some data representations differ, and they use different methods to accom-
plish the same tasks. It’s important to understand that these differences cause problems for
distributed application developers because a company that uses CORBA can’t communicate
with a company that relies on DCOM, and vice versa.

Java RMI
I included Java Remote Method Invocation (RMI) in the chapter because it’s an important
adjunct to CORBA. It isn’t a separate or unique wire protocol because it’s built on CORBA,
but Java RMI does include some features you need to know about. The most important
feature is simplicity. The creators of Java RMI recognized that many developers viewed
CORBA as difficult to use and error prone, so they created an easier to use wire protocol
in the form of Java RMI.
 How SOAP Differs from DCOM and CORBA 21

One of the better overviews of Java RMI appears at http://www.java.sun.com/

marketing/collateral/ rmi_ds.html. I found the diagrams a little difficult to see,
but, otherwise, the information is easy to understand. You’ll also want to view the more
detailed Java RMI white paper at http://www.java.sun.com/marketing/ Ch
collateral/javarmi.html and the specification at http://www.java.sun.com/
products/jdk/1.1/ docs/guide/rmi/spec/rmiTOC.doc.html. 1

All you need to do is spend a little time with Java RMI to realize the simplicity it brings to
the development environment. The protocol automatically takes care of many of the low-
level details you normally need to consider. For example, you can gain access to a remote
object by looking it up in a name facility provided by Java RMI, or by receiving the reference
as a return value or method invocation argument. In addition, the protocol doesn’t need to
consider language or platform differences because communication occurs directly between
two Java Virtual Machines (JVMs). You can see that Java RMI is a simplified form of CORBA
by looking at the block diagram and architectural discussion at http://www.java.sun.com/
products/jdk/1.1/docs/guide/rmi/spec/rmi-arch.doc.html#200.

One thing that might escape your notice at first is that Java RMI also inherits all of the good
features of Java. This means you gain access to niceties like garbage collection. When you
develop an application using CORBA or DCOM, you have to allocate objects and remember
to destroy them manually. Java RMI still requires that you allocate the objects, but the run-
time automatically destroys objects that go out of scope (they’re no longer referenced by
other objects).
The first feature that I noticed beyond simplicity is that the Java RMI doesn’t require a large
support library. You don’t have to install anything extra—everything comes with the JVM.
Many developers I know have had horrifying experiences getting DCOM to work because of
DLL hell. DLL hell is an evil giant of a problem in which older applications either won’t run
with new versions of DLLs or they overwrite those DLLs and cause new applications to fail.
The fact that Java RMI appears as part of the whole Java package and there’s little chance for
incompatibilities to creep in can save more development time than you’d ever imagine.
Java RMI also includes a unique feature—the ability to move behaviors between machines
using inheritance. (Normally, Java RMI extends a default class; but, using special program-
ming techniques allows you to extend other classes as well.) This feature allows a developer
to add a new behavior to an existing class. For example, an existing class may perform finan-
cial analysis, but you might want it to include sales tax or other added calculations. A new
behavior would allow an existing class to perform this task. Contrast this with CORBA and
DCOM, which allow a developer to call methods in existing object or move objects around,
but not augment existing object behavior.
Simplicity comes with a high price in this case. For one thing, you’re still dealing with a
binary protocol. It’s still impossible to get messages past firewalls in some cases. In addition,
you have less control over the environment when problems occur (simplicity gets in the way).
22 Chapter 1 An Overview of Soap

Another problem is that Java RMI is understandably a Java only solution. As long as you’re
willing to make your programming language Java, you’ll be fine. However, Java isn’t noted
for its high execution speed and is, therefore, less suitable than other languages for some
types of desktop and automation applications.
Java RMI passes all parameters by copy rather than by reference. This means a remote
method can’t change the value of an argument you pass; it must provide a return value in
other ways. Passing parameters by copy is the Java RMI method of handling some of the
intricacies of marshalling. Java RMI always passes objects, on the other hand, by reference.
The client gains access to an interface, not an object implementation.

SOAP, HTTP, and XML

I can already hear some of the SOAP savvy among you sighing that I chose to lump SOAP,
HTTP, and XML together. It’s true that these protocols are useful as separate entities and
don’t necessarily depend on each other. However, the real world situation today is that
developers do use them together, so for the sake of simplicity, I’m lumping them together
for now.
So, what does a SOAP message look like? It depends on the technologies that you use with
SOAP, which is why I specified the three technologies I’m using up front. Here’s a simple
example—so simple, in fact, that you probably wouldn’t use this form in a real message.
<SOAP-ENV:Envelope>
<SOAP-ENV:Header>
</SOAP-ENV:Header>
<SOAP-ENV:Body>
<GetPerson>
<LastName>Mueller</LastName>
<FirstName>John</FirstName>
</GetPerson>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>

This example points out several similarities between Web technology that you’re already
familiar with and SOAP. Notice that everything has an opening and closing tag—that’s the
XML influence on SOAP, but also has it’s roots in the HyperText Markup Language
(HTML) we’re all familiar with. A SOAP message always appears within an envelope. The
envelope can have a header, just like an HTML document would, but this isn’t required.
The message must have a body. The message content appears within the body as shown
here. We’re making a request of the GetPerson function for a person whose last name is
Mueller and first name is John.
You can extend all of these tags. In fact, you’ll likely extend most of them when creating
standard SOAP messages. For example, the envelope tag will probably look something like
this in a real message.
<SOAP-ENV:Envelope xmlns:SOAP-ENV=”http://schemas.xmlsoap.org/soap/envelope/”>

The second parameter, in this case xmlns, is the namespace used for the envelope. A
namespace reference tells where to find definitions for classes that contain functions used
 SOAP, HTTP, and XML 23

within the message. So, let’s expand this principle to the GetPerson function mentioned
previously. Here’s what the body of the example message might actually look like.
<SOAP-ENV:Body xmlns:MyObj=”http://www.mycompany.com/myobjects/”>
<MyObj:GetPerson>
<LastName>Mueller</LastName> Ch
<FirstName>John</FirstName>
</MyObj:GetPerson> 1
</SOAP-ENV:Body>

The xmlns tag contains the location of the MyObj object on your company server. This
object contains the GetPerson function used in this example. As you can see, a SOAP
message could quickly begin to look complex, but it’s simple in design.

Remember that you must always pair all XML and SOAP tags; there’s always an opening
and closing tag. In cases in which you don’t need to include any data between tags,
you can signify the closing tag by adding a slash at the end of the tag like this example
of a paragraph tag: “<P />”.

We’ll discuss namespaces in more detail as the book progresses. For now, all you need to
know is that SOAP messages have a specific format that looks familiar if you’ve worked with
XML or even HTML in the past.
OK, I’ve shown you the XML and SOAP part of the picture; so where does HTTP come
into play? SOAP uses HTTP as a transport protocol—the method for getting a message
from one place to another. A SOAP message transaction normally occurs in two phases. The
first appears within an HTTP request message. The client requests data from the server, just
as it would for a Web page. The server places the requested data within the SOAP message
that Web server packs into an HTTP response message. The Web server uses the same tech-
nique as it would normally use—only the content of the HTTP message differs from normal.
A standard browser can view the SOAP message you create. For example, Figure 1.2 shows the
output from the sample code we looked at earlier. Since we didn’t create an XSL file for this
example, the browser uses the default template. This template shows you the content of the
SOAP message with keywords and content highlighted in a variety of colors. Normally, you’d
pass the SOAP response message to a client side application for interpretation and display.

Now that you have a taste of what SOAP is like, you might want to view the SOAP
specification. The current specification, as of this writing, is version 1.1. You’ll find it at
http://static.userland.com/xmlRpcCom/ soap/SOAPv11.htm (among
other places). Not only will the specification fill you in on additional details, but you’ll
also see some example messages that contain the HTTP header, as well as the SOAP
message. I chose to concentrate on just the SOAP portion of the message in my
example listings.
24 Chapter 1 An Overview of Soap

Figure 1.2
Internet Explorer
versions 5.5 and
above can display
SOAP messages in
a simple format.

Problems Solved by Using SOAP

Some developers are dubious about the benefits of using SOAP because they see it as a com-
plex way to transfer data from one location to another using unsafe methods. For example,
SOAP doesn’t have any form of internal security and you can’t use it when you need trans-
actions. One developer even stated that SOAP was a poor choice for any kind of database
work, unless you include safeguards at both ends of the transaction. (This is where all of
those classes on three-tier design come into play—SOAP should access only the business
logic on your server, not the database itself.)
No matter what the developer community comes up with for a data transfer protocol, there
are always going to be negative points to consider. SOAP is a “fire and forget” solution to
transferring data on the Internet that works well in some cases and not at all in others. I’m
not going to convince you in this section that SOAP is the perfect solution for every prob-
lem. You shouldn’t consider it the solution for all problems either—SOAP has limitations.
All arguments aside, SOAP does solve two very important problems and a host of smaller
problems. The first problem is the matter of network security versus data transfer. DCOM
and CORBA are both binary data transfer protocols. This means they both rely on non-
ASCII characters that a Web server firewall could interpret as executable code and packet
formats that the firewall will interpret as invalid HTTP packets. In both cases, the firewall
throws the packets out because it’s designed to work that way. Since the Web server requires
the firewall to keep crackers out, there’s an obvious problem with using binary data transfers.
SOAP solves the problem by using ASCII in an HTTP package—something the firewall
will instantly recognize as data and pass along to the Web server for processing.
 Problems Solved by Using SOAP 25

The second major problem is one of compatibility. If you use a binary protocol, the client
and server must both use the same method for retrieving the data from the packet. Binary
compatibility is easy to achieve when programmers in one company maintain both the client
and server. However, distributed applications aren’t one-company solutions. You now have
Ch
other companies in the form of partners and clients to consider. Transferring ASCII data
within an HTTP packet between companies is easier and less error prone than binary 1
methods. SOAP combined with UDDI makes it easy for customers to find you and use
the resources you provide. Services such as online ordering are easier to implement when
everyone has common grounds for participation.
It’s ironic that one of the features that makes binary protocols so useful on a LAN actually gets
in the way when working in a distributed environment. DCOM provides a wealth of security
setting features that enable you to fine-tune your approach to keeping your data secure. For
example, you can choose to check a user’s identity only once at the outset of a communication.
On the other hand, you can ask the user for identification information before each transaction.
While this does aid in security, it also means that DCOM is doing a lot of handshaking with
the application; a feature low bandwidth connection like the Internet won’t support. SOAP
doesn’t implement any security internally, so there isn’t any security-related handshaking to
worry about when using this protocol. SOAP relies on security like SSL that’s designed for
Internet use, so again, the effect of security on system performance is minimal.
Proxy servers can also present problems for the developer. The target for a remote call
might not have a publicly routable IP address. This means that a proxy server will have to
route the packets to the correct internal address—adding yet another failure point where
incorrect configuration information can keep an application from working. SOAP is
interpreted by a listener component that has a direct public IP address. You still gain
security benefits because the listener doesn’t do anything with the data—an internal
(nonaccessible) component works with the data after the listener interprets it. There’s
plenty of opportunity to check the data for security and corruption problems as part of
the translation process.
Older wire protocols often rely on nonstandard TCP/IP ports for communication so they
don’t interfere with standard network traffic. One of the more common ports is 135, but
these protocols can use other ports as well. What happens if a network administrator recon-
figures the firewall to close the port the application needs? It may take a long time to find
the problem because this is an unexpected error. DCOM and CORBA will simply report
they can’t contact the server—leaving you in the dark as to the cause of the problem. SOAP
solves this problem by using standard communication ports that a network administrator is
unlikely to close.
This is the tip of the iceberg. Microsoft and other vendors designed SOAP primarily to
solve the problems of Internet communication. It isn’t too surprising that it excels in the
kinds of communication that you’d expect to perform on the Internet. You’ll run into
problems, however, if you assume that SOAP is the only answer. DCOM, CORBA, and Java
RMI still provide many useful features not found in SOAP that address distributed applica-
tion programming within a company or tightly knit group of companies.
26 Chapter 1 An Overview of Soap

Performance Issues
It’s important to realize that SOAP represents a new method of transferring data from one
point to another. Unlike binary technologies used with LANs, SOAP travels between dis-
tributed applications on a public network using ordinary text. The differences between
SOAP and other transport methods like DCOM Wire Protocol mean that you can’t rely
on old technique to ensure swift transmission of data.
The following sections will look at the two sides of the SOAP performance coin. On the
one hand, there’s good reason to suspect that SOAP won’t perform well as binary data trans-
fer methods. For one thing, it’s a verbose data transfer method. On the other hand, SOAP
doesn’t require complex translation methods on both ends of the wire. Plain text can get
through firewalls with relative ease and is understandable by all platforms.

Tradeoffs of Using SOAP

There are many performance tradeoffs to consider when working with SOAP. One of the
most obvious performance tradeoffs is the compactness of binary data when compared to
the ASCII text used by SOAP. Although ASCII text is generally compatible with any system
on the market, it’s also bulky. (Look at the simple coding examples in the “SOAP, HTTP,
and XML” section of this chapter for details.) Unfortunately, there isn’t any direct way to
alleviate this problem. You can’t compress the data without losing some of the compatibility
that SOAP is supposed to provide.
The size of your messages is a concern because SOAP works on low bandwidth networks for
the most part. The Internet isn’t as fast as a LAN or other local dedicated connection. In
fact, large message sizes will also affect local server resources (more storage space and longer
handling times). Overall, message size is the one performance problem that you’ll have a
hard time solving.
You can secure a SOAP transaction if you want to. For example, since SOAP relies on
HTTP as a transport protocol, you have access to all of the same security measures that you
can use with HTTP like Secure Sockets Layer (SSL). However, as anyone who has ever
used SSL will tell you, connection speed suffers greatly. It doesn’t take long to realize that
you can either have a slow secure SOAP connection or a fast connection that gives everyone
access to your data. SOAP lacks the alternatives of other protocols such as DCOM provide.
Another potential problem is a two-edged sword. SOAP messages consist of plain ASCII
data in XML format. This means that an application on the client has to translate binary
data into an ASCII format for transport. Another application on the server end has to
translate the data from ASCII format back into a binary format. The translation of data
takes time. Just how much time depends on the amount of data you need to transfer and
the original data format. Transferring graphics or other strictly binary data is an expensive
proposition, while database fields are less problematic if they contain a lot of string data to
begin with. (You’ll notice that many SOAP examples are queries for data from a database,
but few show record updates or transfers of data that’s normally interpreted from a binary
format.) We’ll see in the next section of the chapter that there’s an up side to the use of
ASCII text to transfer your data that could become a performance boost.
 Performance Issues 27

When SOAP Is Faster

The SOAP performance picture isn’t entirely grim. In fact, it can be a faster transport mecha-
nism than those binary alternatives. Consider for a moment that we’re talking about a distrib-
uted application environment with firewalls and other obstacles to overcome when transferring Ch
data from one point to another. A SOAP message requires no translation to get past the
firewalls—binary data may very well require translation, if it gets past the firewall at all. 1
Here’s the second part of the two-edged sword mentioned in the previous section. There’s
no simple answer to the question of whether it takes more time to transfer binary or ASCII
data over the Internet because there are too many variables to consider. Binary data transfers
could take more time if a firewall configuration issue forces the server to resend the data
multiple times. On the other hand, you have the size of the SOAP message to consider. The
only way you’ll know how well SOAP is going to perform is to test it in actual use and make
a comparison to similar tests using technologies such as DCOM. If performance is a major
criterion, then you need to perform such tests.
Another, more subtle, performance difference is that SOAP requires no handshaking. When
you’re working with a low bandwidth, high latency media like the Internet, handshaking can
become a real performance problem. DCOM requires keep-alive messages and other house-
keeping messages that don’t significantly affect the performance of a LAN, but eat precious
bandwidth on most Internet connections.

Sometimes performance tuning an application means getting creative, rather than accept-
ing you can’t do much to improve performance. For example, reducing the length of
string fields can mean the difference between transferring one or two packets.
Reducing the number of packets will increase performance. If you can’t reduce the size of
the strings, then try to use keywords in place of some of the fields. For example, if a field
can only contain a limited number of selections, try passing a number instead of a string.
The component on the receiving end can translate the number back into a string as part of
the process of reading the SOAP message.

SOAP can also create smaller messages in some cases. A binary protocol uses large object refer-
ence tags in some situations. The resulting message is actually larger than the SOAP message.
A DCOM packet contains information other than the data that you're asking the protocol to
transfer, which is the reason for the increase in message size. Admittedly, this is only the case
when the object is relatively small. However, if you're making a lot of small requests, rather
than a few large requests, the difference in request header size can make a difference.

Don’t get the idea that there’s any free lunch when it comes to SOAP. Yes, DCOM uses
a large object reference tag that could adversely affect performance, but DCOM also
supports features like properties. SOAP and DCOM aren’t equivalent; DCOM provides
more functionality than SOAP for the near future, so you may still find that you need
DCOM functionality even if using it does hurt performance.
28 Chapter 1 An Overview of Soap

SOAP and the Web Server

When you’re working with SOAP, you’ll eventually need to deal with a Web server. Since
Web servers differ in capability and features, it’s almost certain that you’ll run into problem
getting SOAP to work with some of them properly. For example, most developers worry
about interoperability between Internet Information Server (IIS) and the Apache server used
by many UNIX servers right now.
You’ll find two common problems working with Web servers from different vendors. First,
there’s the problem of data reception. A Web server commonly uses a listener component
to pick up SOAP messages and act upon them. If every listener acted the same way, there
wouldn’t be a problem. However, there are nuances of difference between SOAP listener ver-
sions, so developers often run into compatibility problems. The amount of trouble you run
into will depend on the SOAP toolkit you use to create a solution. There are some listener
components written in Java that purport to fix these problems, but little idiosyncrasies remain.
Another problem is one of translation. Getting the data to the server doesn’t help much
if the server can’t interpret it. As we saw earlier, SOAP relies on an XML format to get
data from one place to another. What happens if the server is expecting tags in one
order and the client provides them in another? Obviously, a smart developer creates
components that will interact with the data no matter what order it comes in, but
unforeseen problems can occur during the initial design process.

Why Make the Move to SOAP?

SOAP isn’t a magic bullet designed to kill the problems that plague most of us in a world of
ever-increasing complexity. A distributed application requires time and patience to build—
a new technology is unlikely to reduce that complexity to any extent. However, SOAP can
reduce the difficulty of implementing a distributed application solution by reducing the number
of potential problem areas. The mere fact that SOAP makes it easier to move your data
through firewalls is enough to use this technology when creating distributed applications.
Debugging is another reason to use SOAP. When working with DCOM and other binary
technologies, you can’t really see the data flow between machines. You can see the data
before the sender transmits it and you can see it once it arrives at the destination, but you
can’t see the part in-between (unless you’re very good at reading the binary data in packets).
Since SOAP relies on plain text transmission, debugging is easier because you can see flaws
in the data transmission itself.
Developers who create many distributed applications often cite compatibility as the reason to
use SOAP. Using a binary protocol implies that every machine will understand that protocol
equally. In other words, if you choose CORBA IIOP to transfer your data, every machine
involved needs to speak that protocol. (One kludge that developers use is to write code that
bridges translation differences between different protocols such as DCOM and CORBA, but
this is an error-prone method of handling the problem.) SOAP promises to provide a platform
independent transfer media since every computer speaks plain text. The only requirement is
 Case Study 29

that the computers use the same standard to translate the XML text into data formatted for the
application—a much easier requirement to fulfill.
A few developers are finding that a hidden reason to use SOAP is development speed. SOAP
is a simple protocol when compared to DCOM or CORBA. Not only do you have an Ch
opportunity to debug your application with relative ease, but the coding requirements are
easier to understand and less error prone. Developers run into fewer configuration related 1
bugs in their applications and spend more time locating real problems. Once SOAP is stan-
dardized, you may find that the biggest developer specific reason for moving to SOAP is to
get more coding done faster and with fewer errors. Making the developer more efficient
always pays big dividends in productivity and developer happiness.

Case Study
The main reason to use SOAP today is to create distributed applications that run over the
Internet. Yes, you can use SOAP for other types of development and it’ll work fine, but
SOAP isn’t a cure all for every programming problem on your network. It’s one solution
out of many. DCOM and other wire protocols still have a place in your enterprise and you
need to keep that in mind as you develop new applications.
So, how are some companies using SOAP today? Many of the pioneers are still in the testing
stage as I write this. SOAP is an exciting technology, but the standard is still changing on a
daily basis, so many companies have relegated SOAP to research for future products. The
bottom line is that you shouldn’t rush into a SOAP project because it looks interesting.
One company that I talked with recently has begun using SOAP for a mainstream application
and for good reason—SOAP really is the only answer in this case. This company has requested
anonymity, so we’ll refer to them as ABC Corporation for the purposes of this case study. The
fact of the matter is that this company could just as easily be the one that you’re working in now.

Definition of a Problem
The management at ABC Corporation began looking at the company’s bottom line as
profits began to drop in a slowing economy. While ABC Corporation had a great sales staff,
a good product line, and relatively competitive prices, in some cases they couldn’t get their
product out in time to beat competitors. Taking orders by telephone, filling out paper
orders, and the like simply cost the company too much time. Management decided that
direct business-to-business sales over a computer connection would speed things up consid-
erably and reduce labor costs as well. A faster processing cycle could mean the difference
between staying afloat or truly excelling at business.
It’s a nice thought and it’ll probably work. The only problem is that the programming staff
at ABC Corporation had no idea on how to implement the solution. Their company relied
heavily on UNIX servers, while the majority of their customers had Windows NT or
Windows 2000 servers. Direct data communication was impossible—at least direct commu-
nication in the form expected by management. Automation might prove impossible if the
development staff didn’t find a common form of communication.
30 Chapter 1 An Overview of Soap

At first, ABC Corporation tried the old standby of using forms on their Web site and
Common Gateway Interface (CGI) scripts to process the orders. Unfortunately, the installa-
tion proved less robust than management wanted. It was also slow, error prone, and didn’t
consider individual company needs. Although the system might have worked, no one used it
regularly because it was cumbersome and time-consuming—typing out a paper form was
still easier.

Light at the End of the Tunnel

ABC Corporation tried a number of other solutions. All of these solutions failed for one
reason or another. The final straw was when they discovered that the Java RMI solution they
implemented was too slow and cumbersome. Management decided to call in a consultant.
A number of consultants suggested ways to get around the company’s problems, but ABC
Corporation finally decided to try SOAP. They decided that SOAP would circumvent the
majority of the problems they knew about and the other problems were easy enough to
solve using other methods.
Development went slow—a concession to being the first on the block with a new solution.
However, the company eventually had a workable solution to test. Profitability had to be
right around the corner.

Perfection, an Ongoing Process

Life is often a twisted path, rather than a straight road. ABC Corporation found that their
new solution was less than perfect. Compatibility problems between the Windows imple-
mentation of the SOAP listener and the Apache version caused a few hours of nail biting
as the consultant and development staff rushed to fix the problems. The new code worked
around the compatibility problems, performed a few additional checks, and included some
more error trapping.
While the application was now stable, it did execute slightly slower than before. Sometimes
there isn’t a perfect solution, just one that works. The lesson that ABC Corporation learned
in this case is that they should have spent more time looking for a SOAP toolkit that works
equally well on Windows and Apache.
Some of the problems were outside the control of ABC Corporation or their partners. One
of the things that SOAP assumes is a perfect network and we all know that the Internet is far
from perfect. Although ABC Corporation has improved system up time and provided their
partners with alternative means for submitting orders when needed, glitches still occur. They
don’t occur on a regular basis, but even one lost order might result in the loss of a customer.
The lesson learned here is that you need to build plenty of redundancy into your SOAP
implementation and provide alternative communication methods in the case of failure. ABC
Corporation also added an email response to their setup. The customer receives an email
when ABC Corporation receives the order and another message when they send the order
out. This additional communication is completely automated, so it costs ABC Corporation
little and greatly enhances customer satisfaction.
 Case Study 31

ABC Corporation plans to roll out their SOAP solution to all partners in about three months
as I write this. The testing phase is still ongoing. They currently have a pilot program for
several partners and the results look promising. It would be difficult to place a precise date on
return on investment at this point. However, ABC Corporation can process orders faster,
Ch
with less human intervention, and greater customer satisfaction. Just these changes make the
bottom line look a lot better. 1
 CHAPTER
2
SOAP in Theory
In this chapter
Dissecting the SOAP Message 35
Using SOAP with Your Current Code 45
Discovering SOAP Services 47
Putting Everything Together 50
Using SOAP to Move Data 52
Understanding SOAP Attachments 57
Case Study 58
34 Chapter 2 SOAP In Theory

This chapter answers the question “How does it work?” It’s important to know the answers
to this question because you’ll want to troubleshoot your SOAP applications after you begin
developing them. I’ll provide you with details about the parts of SOAP that I know you’ll
need to create applications. The chapter also contains pointers to Web sites that contain
deep theoretical information that’s nice to know, but you really don’t need at them this
moment to create applications. In short, this chapter serves as a practical guide to SOAP
with pointers to the parts that only an engineer could love.
SOAP is a new technology seeking the approval of a standards group. What this means to you,
as a developer, is that SOAP will continue to change shape until the standards committee
finally settles on a set of features. During the next few months you’ll find yourself down-
loading new feature lists, upgrading your tools, and spending a lot of time reading the latest
theory on how SOAP should work. New technology is great because it allows you to
perform tasks that you couldn’t perform in the past, but it also requires extra patience when
things don’t work as well as they should.
The first section of this chapter, “Dissecting the SOAP Message,” tells you how a SOAP
message is put together. We’ll begin looking at the entire package in detail, everything from
the HyperText Transfer Protocol (HTTP) header to the eXtensible Markup Language
(XML) used to create the SOAP message. In the end, you’ll know how a typical message
looks so that you can diagnose problems with messages that your applications create.
The second section, “Using SOAP with Your Current Code,” provides some basics on
retrofitting existing code to use SOAP. Many organizations created DCOM or CORBA
solutions for distributed applications needs in the past. This section provides an overview
of some of the pitfalls of converting these applications to use SOAP.
In the third section, “Discovering SOAP Services,” you’ll learn some of the technologies
available for finding a SOAP service you need for an application. The three technologies
that we’ll discuss are Discovery of Web Services (DISCO), Web Service Description
Language (WSDL), and Universal Discovery, Description, and Integration (UDDI).
Section four will put the final pieces of a basic SOAP application into place. I’ll tell you how
everything works from a theoretical perspective. Of course, theory is often different from
reality, which is why this book has far more hands on chapters than it does theory. We’ll also
visit some of the pitfalls that the hands on chapters help you solve.
Of course, the whole purpose behind using SOAP is to move data. Section five, “Using
SOAP to Move Data,” addresses that issue. It looks at how you deliver data using SOAP.
More importantly, it discusses some of the current limitations on moving data using SOAP
and how the vendors involved in the SOAP effort plan to address that need.
The next section of the chapter, “Understanding SOAP Attachments,” looks at one alterna-
tive for moving data from one point to another. The SOAP attachment is just one of several
alternative delivery methods currently under discussion. Many developers feel that attach-
ments offer the most promise because they’ll allow SOAP to deliver any type of data. In
addition, a developer can encrypt an attachment, improving overall security.
 Dissecting the SOAP Message 35

The final section of the chapter contains a case study. We’ll look at how one company used
SOAP to overcome vendor relationship problems. SOAP works best in an environment
where more than one company needs to exchange information and the data exchange occurs
over a public network like the Internet.

Dissecting the SOAP Message

You can divide SOAP messages into two basic categories: Requests and responses. The
client sends a request to the server. If the server can fulfill the request, then it sends a data
message back to the client. Otherwise, the server sends an error message indicating why it Ch
couldn’t send a response back to the client. In most cases, the problem is one of security,
access, equipment failure, or an inability to find the requested object or data. 2
SOAP messages don’t exist within a vacuum. If you send just a SOAP message, it will never
reach its destination. Remember from Chapter 1 that SOAP is a wire protocol—it relies on
another protocol for transport. This is the same technique used by other wire protocols, so
there’s nothing strange about SOAP when it comes to data transfer needs. The most com-
mon transport protocol in use today is HTTP, so that’s what we’ll look at in this section.
Keep in mind, however, that SOAP can theoretically use any of a number of transport
protocols and probably will in the future.

Why Use Another Transport Protocol for SOAP?

I keep talking about how you shouldn’t assume that HTTP is the only transport protocol you should use with SOAP.
Accepting this statement at face value is one thing, understanding why is another. It’s important to know why you’d
need another transport protocol, other than sheer stubbornness in not liking HTTP for whatever reason.
The first reason is the one that most businesses will eventually accept as the main reason to change to another
protocol. HTTP is a synchronous protocol. In other words, the client and server have to exist at the same time
and generally communicate in real time. Not all business applications fit this model very well, which is why
technologies such as Microsoft’s Queued Components are popular. Your business may eventually use another
transport protocol to gain asynchronous message support.
The second reason is security and privacy concerns. The HTTP protocol is very good at what it’s designed to do.
However, as various countries begin to grapple with the issues of security and privacy, they may find that HTTP
is too open to meet the demands of modern business. While HTTP continues to provide services when you surf
the Web, developers may relegate it to the closet in business communication.
Finally, there are practical reasons to use some other transport protocol. As we’ll see in the “HTTP Extension
Framework” section of this chapter, vendors are currently considering some standard method for showing
differences between HTTP standard elements and extensions to that standard. Remote procedure calls (RPCs)
fall into the extension category—they aren’t a part of the HTTP standard. Using another transport protocol
might become necessary if HTTP extensions become a compatibility problem.

SOAP messages are also a special form of XML. Therefore, in addition to the HTTP
wrapper, a SOAP message requires an XML wrapper. All that the XML wrapper does, in
this case, is tell the data receiver that this is an XML formatted message. The SOAP part
of the message contains all the data; however, SOAP uses XML to format the data.
36 Chapter 2 SOAP In Theory

Figure 2.1 shows a common SOAP message configuration. Notice the SOAP message
formatting. This isn’t the only way to wrap a SOAP message in other protocols, but it’s
the most common method in use today.

Figure 2.1
An illustration of how
a SOAP message is
commonly encased
within other protocols.

The next three sections are going to show you how a SOAP message actually appears
during transmission. We’ll use Figure 2.1 as an aid for discussion. It’s the only time we’ll
explore a complete request or response in the book because you only need to worry about
the SOAP message in most cases. The first section looks at the HTTP portion of SOAP.
The second helps you understand where XML comes into play. Finally, you’ll look at the
SOAP message itself.
 Dissecting the SOAP Message 37

Working with the new capabilities provided by technologies like XML and SOAP means
dealing with dynamically created Web pages. Although it’s nice that you can modify the
content of a Web page as needed for an individual user, it can also be a problem if you
need to troubleshoot the Web page. That’s where a handy little script comes into play.
Type javascript:’<xmp>’+document.all(0).outerHTML+’</xmp>’ the
Address field of Internet Explorer for any dynamically created Web page and you’ll see
the actual HTML for that page. This includes the results of using scripts and other page
construction techniques.

Ch

Viewing the HTTP Portion of SOAP 2

The HTTP portion of a SOAP message looks much the same as any other HTTP
header you may have seen in the past. In fact, if you don’t look carefully, you might pass
it by without paying any attention. Like most HTTP transmissions, there are two types
of headers—one for requests and another for responses. Figure 2.1 shows examples of
both types.
As with any request header, the HTTP portion of a SOAP message contains an action
(POST, in most cases), the HTTP version, a host name, and some content length informa-
tion. The POST action portion of the header contains the path for the SOAP listener, which
is either an ASP script or an ISAPI component. Also located within a request header is a
Content-Type entry of text/xml and a charset entry of utf-8. The utf-8 entry is important
right now because many SOAP toolkits don’t support utf-16 and many other character sets
(at least not as of this writing).
You’ll also find the unique SOAPAction entry in the HTTP request header. It contains
the Uniform Resource Identifier (URI) of the ASP script or ISAPI component used to
parse the SOAP request. If the SOAPAction entry is “”, then the server will use the
HTTP Request-URI entry to locate a listener instead. This is the only SOAP specific
entry in the HTTP header—everything else discussed can appear in any HTTP
formatted message.

The SOAPAction entry currently works with the HTTP protocol, but not other protocols
such as Simple Mail Transfer Protocol (SMTP). Vendors involved in the SOAP specifica-
tion are currently discussing the use of the SOAPAction entry. The SOAPAction entry may
have changed a little by the time you read this, so you can use it with other protocols.

An optional request header entry is the mandatory HTTP Extension Framework declaration.
You’ll notice that the action for such a request header is different. For example, instead of
POST, you’ll see M-POST in the header. This type of header also includes an entry such as
Man: “http://schemas.xmlsoap.org/soap/envelope/”. The effects of the HTTP Extension
Framework are discussed later in the chapter.
38 Chapter 2 SOAP In Theory

UTF stands for Unicode Transformation Format. UTF represents one-way standard
method for encoding character. One of the better places to learn about UTF-8 is
http://www.utf8.org/. You can find a good discussion of various encoding tech-
niques at http://www.czyborra.com/utf/. This Web site presents the informa-
tion in tutorial format. The fact remains that you need to use the UTF-8 character set
when working with SOAP.

The response header portion of the HTTP wrapper for a SOAP message contains all of the
essentials as well. You’ll find the HTTP version, status, and content length as usual. Like
the request header, the response header has a Content-Type entry of text/xml and a charset
entry of UTF-8.
There are two common status indicators for a response header: 200 OK or 500 Internal
Server Error. While the SOAP specification allows leeway in the positive response status
number (any value in the 200 series), a server must return a status value of 500 for SOAP
errors to indicate a server error.
Whenever a SOAP response header contains an error status, the SOAP message must
include a SOAP fault section. We’ll talk about SOAP faults later in this chapter. All you
need to know now is that the HTTP header provides the first indication of a SOAP fault
that requires additional processing.
There are other applicable status error codes for the response header. For example, if the
client sends a standard HTTP header and the server wants to use the HTTP Extension
Framework, it can respond with a status error value of 510 Not Extended. The 510 error
isn’t necessarily fatal; a client can make the request again using the mandatory HTTP
Extension Framework declaration. In this case, an error message serves to alert the client
to a special server requirement.

Viewing the XML Portion of SOAP

All SOAP messages are encoded using XML. SOAP follows the XML specification and can
be considered a true superset of XML. In other words, it adds to the functionality already
in place within XML. Anyone familiar with XML will feel comfortable with SOAP at the
outset—all you really need to know is the SOAP nuances.
Although the examples in the SOAP specification don’t show an XML connection (other
than the formatting of the SOAP message), most real world examples will contain at least
one line of XML specific information. Here’s one example of an XML entry:
<?xml version=”1.0” encoding=”UTF-8” standalone=”no”?>

As you can see, the tag is quite simple. The only bits of information that it includes are the
XML version number, the character set (encoding), and whether the message is standalone.
As with the HTTP header, you’ll need to use the UTF-8 character set for right now. The
version number will change as new versions of XML appear on the scene. The standalone
attribute determines if external markup declarations could affect the manner in which this
 Dissecting the SOAP Message 39

XML document is processed. A value of no means external documents could affect the
processing of this document.

We won’t discuss all the XML tag attributes (declarations) in this chapter. You can find a
complete listing of these attributes at http://www.w3.org/TR/REC-xml. For those of
you who don’t read specifications very well (or prefer not to), look at Tim Bray’s anno-
tated XML specification Web site at http://www.xml.com/axml/testaxml.htm.
Another good place to look is the XML.com Web site at http://www.xml.com/.
Finally, if you want to see the tools and other resources available for XML, look at
http://www.projectcool.com/ developer/xmlz/xmlref/examples.html.
Ch

2
Some developers don’t include all of the XML tag attributes in their SOAP messages. So
far, I haven’t seen any problems with leaving the encoding and standalone attributes out of
the picture. You should, however, always include the XML version number—if for no
other reason than the need to document your code and ensure there are no compatibility
problems with future SOAP implementations.

Working with the SOAP Message

You’ve already seen several examples of simple SOAP messages. A message consists of an
envelope that contains both a header and a body. The header can contain information that
isn’t associated with the data itself. For example, the header commonly contains a transac-
tion ID when the application needs one to identify a particular SOAP message. The body
contains the data in XML format. If an error occurs, the body will contain fault informa-
tion, rather than data. Chapter 1 shows a simple example of a SOAP message, along with a
detailed description of the various message elements.
Now that you have a summary of the SOAP message content, let’s look at some particulars
you’ll need when working with SOAP. The following sections will explain some technical
details needed to understand the SOAP message fully. The first section tells you about the
SOAP data transfer requirements as opposed to those supported by HTTP. The second sec-
tion tells you about a Web site that allows you to test your SOAP knowledge. The third
section explores SOAP fault messages. Finally, the fourth section helps you understand the
HTTP Extension Framework.

HTTP and the SOAP Transfer

SOAP is essentially a one-way data transfer protocol. While SOAP messages often follow a
request/response pattern, the messages themselves are individual entities. They aren’t linked
in any way. This means that a SOAP message is standalone—it doesn’t rely on the immediate
presence of a server, nor is a response expected when a request message contains all of the
required information. For example, some types of data entry may not require a response
since the user is inputting information and may not care about a response.
40 Chapter 2 SOAP In Theory

The envelope in which a SOAP message travels, however, may provide more than just a
one-way transfer path. For example, when a developer encases a SOAP message within an
HTTP envelope the request and response both use the same connection. The connection is
created and maintained by HTTP, not by SOAP. Consequently, the connection follows the
HTTP way of performing data transfer—using the same techniques as a browser to request
Web pages for display.

Testing Your SOAP Knowledge

Microsoft recently made two Web sites available to check your SOAP messages. The first
accepts SOAP messages, parses them, and provides a check of their validity. You’ll find it at
http://www.soaptoolkit.com/soapvalidator/. Figure 2.2 shows what this Web site looks like.

Figure 2.2
The SOAP Message
Validation site allows
you to test your SOAP
knowledge and learn
about specific
message types.

As you can see, there are three panes in this display. The SOAP Message Text window
allows you to enter a message that you want to verify. You can also choose one of the valid
or invalid samples from the drop down list boxes. These samples can teach you quite a bit
about SOAP. They provide examples of what you can and can’t do within a message and the
results that you’ll get when performing certain actions. I’ve actually learned how to distin-
guish some error messages by using this Web site. You don’t need to include the HTTP or
XML headers in the SOAP Message Text window, just the SOAP message.
 Dissecting the SOAP Message 41

The Parsed Version window shows what the message looks like after the SOAP listener
parses it. This window doesn’t tell you about the validity of the message, but it does help
you understand the XML formatting better. You can use this window to determine whether
the message is well formed. The tags should form pairs that are easy to see in the pane. The
use of text coloring also helps you to distinguish between specific text elements.
The Message Check Results window shows the results of diagnostics the site performs on
your SOAP message. You’ll see error messages in places where the SOAP message doesn’t
contain required entries or the entry format is incorrect. When all of the error messages are
gone, the SOAP message is ready for use. Of course, this still doesn’t mean the SOAP mes-
sage will do anything. The Web site doesn’t check the validity of the data within the SOAP Ch
message. You can create a perfect SOAP message that still doesn’t work because the server-
side component is expecting the data in a different format or even requires other arguments.
2
The second Web site is a generic SOAP listener. You can send a SOAP message to it using an
application. The site will test the message much like the Message Check Results window of
the SOAP Message Validation site. You’ll find this Web site at http://www.soaptoolkit.com/
soapvalidator/listener.asp. Figure 2.3 shows an error output from this site. This is also
an example of the SOAP fault message that we’ll discuss in the next section of the chapter.

Figure 2.3
The SOAP Message
Listener site validates
requests from your
applications.

SOAP Fault Messages

Sometimes a SOAP request generates a fault message instead of the anticipated reply. The
server may not have the means to answer your request, the request you generated may be
incomplete, or bad communication may prevent your message from arriving in the same
state that you sent it. There are many reasons that you might receive a SOAP fault message.
However, you can generally categorize them into four general areas as shown in Table 2.1.
42 Chapter 2 SOAP In Theory

Table 2.1 General SOAP Fault Message Classifications

Message Type Description
Client The client generated a request that the server couldn’t process for some reason.
The most common problem is that the XML format of the SOAP message is
incorrect (malformed). Another common problem is the server can’t find the
requested component or the client doesn’t provide all of the information the
server needs to process the message. If the client receives an error message of
this type, it must recreate the message and fix the problems of the original. The
server usually provides amplifying information if it can determine how the client
request is malformed or otherwise in error.
MustUnderstand This error only occurs when you set the SOAP message mustUnderstand
attribute to 1. The error occurs when the client sends a request that the server
can’t understand or obey for some reason. The server may not understand a
request when the client relies on capabilities of a version of SOAP that the
server’s listener doesn’t provide. A server may not obey a client request due to
security or other concerns. An upgrade of the listener or server-side compo-
nents usually helps in this situation.
Server The server couldn’t process the message even though the request is valid. A
server error can occur for a number of reasons. For example, the server could
run out of resources for initializing the requested component. In a complex
application environment, the server may rely on the processing of another
server that’s offline or otherwise inaccessible. The client should definitely resub-
mit the same request later, when this error occurs. The server usually provides
amplifying information if it can determine the precise server-side error that
occurred.
VersionMismatch SOAP doesn’t have a versioning system. It does rely on the name spaces that you
define to perform useful work. This error message occurs when the SOAP enve-
lope namespace is either incorrect or missing. As of this writing, you should use a
SOAP envelope name space that points to “http://schemas.xmlsoap.org/
soap/envelope/” as shown in the source code examples in this chapter.

It’s possible to create other fault categories for SOAP messages. The list in Table 2.2
conforms to the SOAP specification. The only requirement for SOAP fault categories is
that they follow the formatting requirements for XML namespaces. A good place to
look for additional information about XML namespace formatting is the Namespace in
XML document at http://www.w3.org/TR/REC-xml-names/.

When a server returns a fault message, it doesn’t return any data. Look at Figure 2.3 and
you’ll see a typical client fault message. Notice the message contains only fault information.
With this in mind, the client-side components you create must be prepared to parse SOAP
fault messages and return the information to the calling application in such a way that the
user will understand the meaning of the fault.
 Dissecting the SOAP Message 43

Figure 2.3 shows the standard presentation of a SOAP fault message. Notice that the fault
envelope resides within the body of the SOAP message. A fault envelope generally contains
a faultcode and faultstring element that tells you which error occurred. All of the other
SOAP fault message elements are optional. Table 2.2 provides a list of these elements and
tells you how they’re used.

Table 2.2 SOAP Fault Message Elements

Element Description
faultcode The faultcode contains the name of the error that occurred. It can use a dot
syntax to define a more precise error code. The faultcode will always begin with Ch
one of the classifications listed in Table 2.1. For example, the faultcode in Figure
2.3 consists of a Client error code followed by an XMLERR subcode. This tells 2
you that the request message is malformed because the XML formatting is incor-
rect. Because it’s possible to create a list of standard SOAP faultcodes, you can use
them directly for processing purposes.
faultstring This is a human readable form of the error specified by the faultcode entry. This
string should follow the same format as HTTP error strings. You can learn more
about HTTP error strings by reading the HTTP specification at http://www.
normos.org/ietf/rfc/rfc2616.txt. A good general rule to follow is to make
the faultstring entry short and easy to understand.
faultactor This element points to the source of a fault in a SOAP transaction. It contains a
Uniform Resource Identifier (URI) similar to the one used for determining the
destination of the header entry. According to the specification, you must include
this element if the application that generates the fault message isn’t the ultimate
destination for the SOAP message.
detail You’ll use this element to hold detailed information about a fault when available.
For example, this is the element that you would use to hold server-side compo-
nent return values. This element is SOAP message body specific, which means
you can’t use it to detail errors that occur in other areas, such as the SOAP mes-
sage header. A detail entry acts as an envelope for storing detail sub-elements.
Each sub-element includes a tag containing namespace information and a string
containing error message information.

You can use the presence or absence of various fault message elements to your advan-
tage when determining the cause of an error. For example, if the faultactor element
appears within the fault message, it’s a good bet that the message passed between
multiple servers. In some cases, this means that you can eliminate gross formatting
errors from your list of things to check because several machines parsed the message
without generating an error. The absence of the detail element, on the other hand, may
indicate that the error happened before the component on the server processed the
request message. This means there’s a problem with processing the message, rather
than a problem with the server-side component’s handling of message data.
44 Chapter 2 SOAP In Theory

Let’s look more closely at the detail element. As stated, the detail element doesn’t stand
alone—it acts as an envelope for additional information. Here’s a typical example of a detail
element entry.
<detail>
<e:additfault xmlns:e=”http://www.mycompany.com/myobjects/”>
<message>Couldn’t Create Object</message>
<errorcode>429</errorcode>
</e:additfault>
</detail>

Notice that detail information begins with a namespace entry. In this case, a component
called AddIt that appears in the MyObjects directory on the server had an error. The error
message is that it couldn’t create a required object, while the associated error number is 429.
The client side SOAP message parser could pass this information to the client application
for additional processing. At least the user would have a better idea of what happened and
could alert a network administrator to the problem.
Of course, the detail element isn’t limited to a single component entry. Every component
that has an error could make an entry. For that matter, you might include a single namespace
entry for every error that a single component experiences. A component might actually
provide several error messages that state an entire list of problems that it experienced in the
processing of the SOAP message. For example, rather than return a single message at a time
for faulty arguments, the component could parse all of the arguments and return a list of
errors that covers all of them.

HTTP Extension Framework

Let’s talk for a few seconds about mandatory HTTP Extension Frameworks. The World
Wide Web Consortium (W3C) defines many aspects of the HTTP header. For example, it
defines what the Content-Type declaration contains. These are standardized portions of the
HTTP header that every browser and browser-like application can understand.
Vendors have extended the HTTP header in ways that the W3C never anticipated. For
example, there are extensions for remote procedure call (RPC) mechanisms. SOAP happens
to fall into this class of extensions, so it’s possible that you’ll need to define an HTTP
Extension Framework for your application. An HTTP Extension Framework describes
which extensions the document contains. It also determines how the recipient should
handle them. You can read more about the HTTP Extension Framework in general at
http://www.w3.org/Protocols/HTTP/ietf-http-ext/ and http://www.normos.org/ietf/
rfc/rfc2774.txt.

SOAP uses HTTP Extension Frameworks as a way to define how it will extend a standard
HTTP message. The HTTP Extension Framework includes three additional entries. First,
you’ll add an M- to the beginning of the action portion of the HTTP header. For example,
POST becomes M-POST. Second, you need to add the HTTP Extension Framework iden-
tification we talked about earlier, Man: “http://schemas.xmlsoap.org/soap/envelope/”;
ns=NameSpace. This tells the recipient where to find information about the HTTP Extension
 Using SOAP with Your Current Code 45

Framework, and how to deal with it. The ns declaration tells which namespace to use with this
HTTP Extension Framework. Finally, you need to add a namespace to the SOAPAction entry
like this, NameSpace-SOAPAction. This qualifies the particular action you need to perform.

Using SOAP with Your Current Code

Some developers are like kids in at a candy store. Simply because a technology is new and
exciting, because it’s accepted by their peers as a good technology, all development must use
that technology. SOAP isn’t the kind of technology that lends itself to this perspective. Look
at SOAP as more of a value-added technology. It’s designed to overcome problems with Ch
older technologies and to simplify the programming environment in some cases. (Just how
much simpler remains to be seen—vendors are currently discussing more and more added 2
functionality, which also increases complexity.)
SOAP is a good problem solving protocol, especially for distributed applications that rely
on the Internet. It allows developers to solve communication and interoperability problems
with DCOM and CORBA applications. The following list provides an overview of some of
the applications that you may want to retrofit to use SOAP in an effort to improve reliability
and reduce complexity.
■ Internet Only Applications: One situation where you may want to commit to a SOAP-
only solution is when retrofitting applications that only access the Internet. The benefit
of going this route is increased interoperability, reduced complexity, and fewer prob-
lems with obstacles like firewalls.
■ Applications with Partner Access: Most businesses have partners to work with today.
The partners may require access to company data. Using SOAP to create the partner
access portion of the application makes sense. You may, however, want to continue
using DCOM or CORBA for LAN access because they provide a higher level of
security and you don’t need to worry about integration.
■ Employees on the Road: Using SOAP to service employees on the road is a tossup if you
already have an application in place. Since you have direct control over the employee’s
machine, using DCOM or CORBA isn’t as much of an issue. In fact, you may want to use
something like COM+ with Queued Components, in this case, because this solution actually
provides better functionality than SOAP. However, you may find that getting through the
firewall is a problem with all of these binary solutions and that SOAP is your only option.

For the purposes of this book, the term cracker will always refer to an individual that’s
breaking into a system on an unauthorized basis. This includes any form of illegal activity
on the system. On the other hand, a hacker will refer to someone who performs low-level
system activities, including testing system security. In some cases, you need to employ the
services of a good hacker to test the security measures you have in place, or suffer the
consequences of a break-in. This book will use the term hacker to refer to someone who
performs these legal forms of service.
46 Chapter 2 SOAP In Theory

■ Monitoring Programs: SOAP is a two-edge sword when it comes to monitoring pro-

grams. On one hand, monitoring programs normally deal with data in text format, so
SOAP is a perfect solution. On the other hand, SOAP lacks the strong security of
binary wire protocols, which means that you’re potentially opening the heart of your
system to the prying eyes of crackers. You’ll need to consider the sensitivity of the data
that you’re monitoring when making an application conversion choice in this situation.
■ Satellite Office Applications: Whether SOAP makes sense in this situation depends on
your network setup. If each satellite office has a separate server, then you could maintain
the high security provided by DCOM and CORBA locally. Use SOAP only for server
updates. This solution maintains the good features of the current application implemen-
tation, while adding the flexibility of SOAP to the Internet portion of the data stream.

It’s tempting to think that you could use SOAP in place of DCOM Wire Protocol or Internet
Inter-ORB Protocol (IIOP). SOAP isn’t a direct replacement for binary technologies because
it lacks certain features. We’ll spend more time talking about the problems SOAP faces today
in the “Current SOAP Implementation Problems” section of this chapter. These problems
will likely decrease as the SOAP specification matures. However, there are certain problems
that won’t go away and they affect your ability to use SOAP with your current applications.
One of the biggest problems is that SOAP relies on text data transfers. Sure, text is the rea-
son that SOAP can work with diverse operating system platforms. Using text also means
that you can’t transfer objects, graphics, or other binary formatted data very well. The world
does run on the content of database management systems (DBMSs), but it also needs these
binary formats. We’ll discuss a possible solution in the form of SOAP attachments later in
the chapter. However, even attachments will prove an error prone solution to the problem.
Security is another problem with current SOAP implementations. A common solution today
is to use Secure Socket Layer (SSL) technology to make SOAP more secure. (We’ll discuss
SSL in the “Secure Socket Layer (SSL)” section of Chapter 3.) The problem with this
solution is that it relies on both parties supporting SSL. In addition, you’re now looking at
another binary implementation of what’s supposed to be a text-base protocol. You might
find there’s a price to pay in the form of interoperability and compatibility when adding
security to a SOAP solution. In some cases, you may actually find that DCOM is ahead in
this regard because it provides in-depth security at only a slightly higher performance cost
and with fewer interaction problems after you get it set up correctly.
SOAP also provides limited data type support. This means that some applications you’re using
today will break when you try to move them to SOAP. If your application uses complex data
formats, you might actually be ahead by using one of the binary protocols or figuring out a way
to reduce the complexity of the application’s data model. Theoretically, SOAP will eventually
support complex data types, including objects. The reality today is that SOAP toolkits provide
a bare minimum of data type support.
 Discovering SOAP Services 47

Discovering SOAP Services

Writing a SOAP application isn’t the end of the story. Creating a server-side listener means
that you have the capability of servicing client-side requests. The client still has to find your
service and ask how to use it. When using SOAP within a single organization, discovering
a service is relatively easy. However, what happens when you need to service requests from
third parties? Will a client be able to find your service if they don’t even know your company
exists? The short answer is no.
Discovery protocols provide a means for you to advertise your service in a neutral manner.
What this means is that a client can find your service, request how to use it, and then make a Ch
request from you directly, all without any prior calls to your company. This is the perfect way to
make business-to-business transactions work. An office supply company only needs to advertise 2
its services on a central Web site to make those services accessible to another company.
Obviously, a company could advertise services in many ways. For SOAP to work as a business-
to-business transaction aid, companies have to agree on a standard advertising format. At
the time of this writing, there’s a lot of agreement between small groups, which means
there’s more than one way to advertise your services. Considering SOAP is a new technology,
the plethora of advertising techniques is expected. Eventually, one or two of these advertising
methods will win out and we’ll all use the same technique for broadcasting service informa-
tion. In the mean time, I’ve chosen what I feel are the best advertising methods for this
book at the time of writing.

Although we’re concentrating on three discovery services in this chapter, there are
many others from which to choose. For example, the Electronic Business eXtensible
Markup Language (ebXML) initiative is taking off as I write this. There isn’t enough
information available yet to write much about, but there probably will be shortly. Like
most discovery protocols, this one is designed to promote business-to-business com-
munication. The two interesting bits about this protocol is that it’s supported by the
United Nations and it was designed from the ground up for international use. You’ll
want to keep your eyes on the main ebXML Web site at http://www.ebxml.org/ if
this discovery language interests you.

Eventually you need to decide which individual or set of discovery protocols to use. The
following sections look at three service description protocols:
■ Discovery of Web Services (DISCO)
■ Web Service Description Language (WSDL)
■ Universal Discovery, Description, and Integration (UDDI)

These are the three best-supported and most generic protocols at the time of writing.
Given the volatile nature of SOAP at the moment, however, you’ll want to keep your eyes
open for other protocols that might meet your business needs better.
48 Chapter 2 SOAP In Theory

Understanding Discovery of Web Services (DISCO)

DISCO is a service designed to make it easier to locate and use SOAP services. This particular
service is SOAP specific and a single vendor, Microsoft, currently supports it. In other words,
this may not be the route to go if you need to support solutions from more than one vendor.
The DISCO service relies on a special protocol named SOAP Control Language (SCL) to
allow the discovery of services by remote computers. Someone interested in locating SOAP
resources will contact a DISCO server, download SCL documents of interest, and then use
those documents to refine service selection. Microsoft states that a developer can use SCL
to discover any resource, not just Web services, but the entire technology does seem
directed at public Web services.

DISCO isn’t a complete specification, so you’ll see changes in the coming months.
Microsoft will likely continue working on this specification and introduce it at
some point as an enabling technology for the .NET platform. You can learn more
about DISCO at http://msdn.microsoft.com/xml/general/disco.asp
and http://msdn.microsoft.com/library/dotnet/cpguide/
cpconenablingdiscoveryforwebservice.htm. The current document is more
theory than practical information, but it does give you a good overview of what DISCO
intends to accomplish.

The discovery sequence relies on an algorithm based on standard Web inquiries. A client
uploads a request document to a server, which returns a discovery (.DISCO) document the
client. Depending on the content of this discovery document, the client can either request a
desired resource, or make additional discovery requests.
After a requestor finds a desired discovery document, they’ll read the SCL statements it
contains. Essentially, these statements point to schema definition language (SDL) files
that contain a description of the target component. The application uses the information
obtained to create an instance of the component using SOAP and work with it in the
normal fashion.

Working with the Web Service Description Language (WSDL)

WSDL is a joint effort by several vendors including Ariba, IBM, and Microsoft. It’s an
XML-based method of describing network services. The specification describes a Web
service as an endpoint—the final destination for a client request. The document your appli-
cation receives contains a schema for the requested component on the remote machine. An
application uses the schema to make remote requests.
The vendors involved with creating WSDL designed it to provide complete flexibility and
extensibility. The only implementations available right now use HTTP and SOAP to make
requests. However, future WSDL implementations could use other transport and wire
protocols. This Web discovery language specification also allows use of MIME, which
means that you can make multiple requests in a single message.
 Discovering SOAP Services 49

WSDL, like many of the technologies discussed in this chapter, is a work in progress.
It’s important to keep track of changes to the specification. You can learn more about
WSDL at http://msdn.microsoft.com/ xml/general/wsdl.asp.

When a client makes a request for information from a server, it receives a complex XML
document in return. The document contains the name of the requested component, some
links for finding documents associated with the component, and a full description of the
component’s schema. The document also contains other information, but these are the
important pieces for this discussion. Ch

The document links are actually namespaces. These namespaces define various document ele- 2
ments, just as the namespaces in a regular SOAP document do. Most of these namespaces point
to locations on the server. Of course, there are some of the usual suspects, like xmlns:soap=
”http://schemas.xmlsoap.org/wsdl/soap/”, which define the location of the SOAP schema.

The schema sections include elements for the various method names, expected parameters, and
the parameter types. This means that you know exactly how to access the component at the
outset, rather than experiencing some level of trial and error in discovering the component.
The pros of this discovery technology are that it provides detailed information, great flexi-
bility, and multiple vendor support. Because of these features, WSDL is the best technology
to use for private Web service discovery. You’d use this technology with business partners
and others who need access to your network, but in a situation in which you wanted to exer-
cise some control over the people who gain access to the information. Some of the cons
include a high level of complexity and the lack of a framework right now. Hopefully, the
vendors involved with this effort will correct some of these failings in the near future.

Understanding Universal Discovery, Description, and Integration (UDDI)

UDDI promises to become the discovery language used by everyone involved in building
distributed applications—or at least it’ll grab the majority share of users. The reason is
simple: UDDI abstracts the process of discovering the Web services that another company
offers. It sits on top of the protocols that are already in place and builds on them. In short, it
doesn’t reinvent the wheel the way many Web technologies seem to do today. In addition,
vendors designed this technology for public use so that companies no longer need to know
anything about each other before they discover the services that they offer.
The element that separates UDDI from other Web service discovery technologies is the
registry, which UDDI participants rely on. UDDI replicates the registry information across
all participating servers so companies in diverse locations can find each other. Instead of
having to know which company-specific server to contact for information, UDDI relies on
a central repository of information. It uses a single place to hold information about the Web
servers represented by all companies that use the registry. Someone who wants to know
about a service you provide doesn’t even need to know that your company exists until they
look for the required service. The registration is an XML file that contains these three
categories of information:
50 Chapter 2 SOAP In Theory

■ Contact information, including company name and address.

■ Taxonomic classifications of the business, which includes a listing of the industries that
the company can assist.
■ Technical listing of the Web services that a company offers (or at least those that it
wants to make public).

The vendors who designed UDDI built it on the work done for WSDL. As a result, UDDI
looks surprisingly similar to WSDL at a low level. Of course, UDDI also contains elements
of both XML and SOAP and it relies on HTTP as a standard transport protocol.

There are many great places to find information about UDDI. One of the better places to
find general information is The XML Cover Pages (http://www.oasis-open.org/
cover/uddi.html). You can find a UDDI technology overview at http://www.
oasis-open.org/cover/UDDI_Technical_White_Paper.pdf. The general UDDI
Web page is at http://www.uddi.org/. It provides links for the actual UDDI specifi-
cation, among other things. Microsoft’s UDDI registry announcement appears at
http://www.microsoft.com/presspass/press/2000/Nov00/UDDIBetaPR.
asp. Even though their registry is essentially for testing purposes now, you can still see
how the technology will eventually work. Finally, the Microsoft UDDI SDK download site
is at http://msdn.microsoft.com/library/techart/Progguide.htm.

UDDI isn’t a full-fledged discovery service. It’s designed to allow businesses to find each
other and learn about services they offer. However, UDDI stops at this point. If you want
to learn the details about a particular service, you need to contact the business in question
through its server. UDDI also avoids providing business details such as pricing information,
time required to provide a service, and the geographic service limits for a target company.
The vendors working on UDDI are designing it to fulfill an extended role. For the pur-
poses of this book, UDDI is a technology you can use to learn about Web services provided
by other companies. However, it might help to keep UDDI’s bigger role in mind. Vendors
eventually want to use it to track products and general services as well, making UDDI a
type of electronic yellow pages for the entire world.

Putting Everything Together

At this point, you have some idea of how SOAP works, but may not really know how every-
thing fits together. SOAP follows a logical data flow that includes an interaction of some
type between a client and a server. It relies on a text file formatted using the XML specifica-
tion and includes special entries that describe the content of the SOAP message. Here are
the steps in a typical exchange between client and server.
1. The client instantiates a local SOAP object. It describes the location where the SOAP
message should go, serializes the content that it wants to send, and includes any
special attributes.
 Putting Everything Together 51

2. The local object encapsulates the SOAP message within a request HTTP message.
It then sends this message to the server.
3. The listener on the server receives the message sent by the client.
4. A parsing object decodes the SOAP message. If the message is valid, it instantiates a
copy of the requested object and feeds it the data sent by the client. Otherwise, the
parser sends an error message back to the client.
5. At this point, if the server has everything it needs and the client doesn’t require any
feedback (or is incapable of receiving it), the parser frees the object it created and waits
for another request.
Ch
6. If the client requires feedback, the server generates a SOAP message of its own. This is
the normal course of action when using the HTTP transport, even if the server doesn’t 2
really have anything to say.
7. The Web server encases the SOAP message within a reply HTTP message. It sends
this message back to the client.
8. The client checks the validity of the HTTP message first. If the Web server reports an
error, then the client might have to send the request message again using different param-
eters. For example, the server might require the client to use an HTTP Extension
Framework as described earlier in the chapter.
9. A client-side parser checks the content of the SOAP message. It removes any result
values and sends them to the client. Since the client made the request as an object call,
the parser uses standard object value return techniques.
10. The client reacts to the data returned from the server. In some cases, this may simply
mean doing nothing at all for a successful transfer. In other cases, the client may display
data or perform other tasks. SOAP isn’t some type of mystery protocol—the end result
is much the same as binary protocols such as DCOM.

The learning curve for using SOAP isn’t a matter of understanding how to perform a task in
a new way, but simply using new techniques to perform that task. You can divide a problem
into these five main areas:
■ Modify client applications so they call a SOAP component, rather than create a local
object. This is similar to the task you’d perform for a DCOM or a CORBA application—
only the actual calling syntax differs.
■ Create a client-side component that generates and interprets SOAP messages based on
the input received by the client or server. In this case, SOAP is slightly more complex
than older technologies like DCOM because vendors haven’t introduced all of the
required tools. However, after new technologies like Microsoft’s .NET Framework
appear, the task of creating the client-side component will become easier.
■ Install or create a server-side listener. In many cases, you can use the default listener
that comes with a vendor supplied SOAP toolkit.
■ Create a server-side component to parse the incoming SOAP message. This component
is also responsible for interacting with the data component—the one that will produce
52 Chapter 2 SOAP In Theory

an output. In some cases, this component might also need to perform data translation.
For example, a database might support data types that SOAP doesn’t support directly.
■ Check server-side business logic and data access components to ensure they’ll interact
with SOAP in a safe manner. In some cases, this means setting new policies. You may
find that security concerns affect how you transfer data to the client. In other cases,
you’ll need to change component logic so that the component doesn’t rely on client
responses. The Internet is an unreliable media, so your application needs to allow for
connectivity problems.

You’ll get a better idea of how these five tasks come into play as the book progresses. The
important thing to consider now is that SOAP doesn’t have to present major challenges to
your organization or make big changes in the way you perform tasks. It will make certain
tasks a lot easier to perform and you’ll find that you have flexibility you didn’t have in the
past. The focal point isn’t in the way you’ll perform these tasks, but in the technique you
use to perform them.

Using SOAP to Move Data

The whole reason to use any protocol is to move data. The applications you create move
data requests from a client to a server, and then the reply from the server to the client. A
server might have to call on the services of another server to answer client requests. Two
servers might exchange data in an effort to optimize operations or perform tasks like user
validation. In the end, it all amounts to moving data from one place to another, shuffling it
around a bit to produce a result, and storing it in yet another place.
Therefore, a true measure of a protocol is how well it moves data around. Not only do you
need to consider the speed at which the protocol moves the data, but also the accuracy, effi-
ciency, and reliability with which the protocol moves the data. After all, no one cares how fast
you can move data if you can’t move the data accurately (without any errors). A developer is
also concerned about the ease of moving the data, as well as the flexibility of the data protocol.
So, how does SOAP stack up? You need to ask this question before committing resources to
a solution that won’t work. The first section that follows looks at some of the current imple-
mentation problems with SOAP. These problems won’t stay around forever because SOAP
is an evolving standard. In addition, even though the problems exist today, vendors are
already talking about solutions they plan to implement in the future. Consider this first
section a snapshot of SOAP today. Make sure you spend some time looking at tools that
become available by the time you read this.
The computer industry is also working on other solutions that mimic the capabilities of
SOAP. The two most popular solutions are XML-RPC and XML Protocol (XMLP). XML-
RPC actually appeared on the scene before SOAP and is currently more mature. It allows
you to do some things that SOAP won’t currently allow, but that flexibility has some costs
that you need to know about. XMLP is a new protocol. At the time of this writing, XMLP
 Using SOAP to Move Data 53

is barely out of the proposal stage. We’ll talk about both of these alternatives so that you
have the full story before embarking on your next development project.

In the confusing world of computer acronyms, the W3C committee in charge of XML
Protocol recently changed the acronym from XP to XMLP. Whether this is in deference
to Microsoft Windows XP or not isn’t clear. However, for those of you who were
already familiar with the XP acronym, be sure to look for XML Protocol under the
XMLP acronym from now on.

Ch
Current SOAP Implementation Problems 2
SOAP is a new technology and the specification is still under review as I write this. Microsoft
and other vendors are creating new toolkits that permit developers to create SOAP applica-
tions with a little more ease than writing everything by hand. Of course, with the SOAP
specification and associated toolkits in a state of flux as they are now, it’s only reasonable to
assume that you’ll run into some problems with the various vendor offerings.
Some SOAP problems result from an immature specification. One issue that vendors are
working on is the idea of a unique identifier for SOAP. For example, when you work with
DCOM, you can access a component using a globally unique identifier (GUID). This
means that if two components have the same name, you can still identify them using their
respective GUIDs.
Many developers complain that SOAP lacks resource management tools. For example, SOAP
doesn’t include the concept of connection pooling. The lack of resource management means
that SOAP has to create resources from scratch for every user request, making SOAP ineffi-
cient when compared to other new technologies such as COM+.
SOAP toolkits also have different ways of performing the same task. Let’s look at a con-
crete example of the problems with the current implementations. I used two toolkits and
fed them the same SOAP request. This first listing is what the Microsoft SOAP Toolkit
Version 2 output.
<?xml version=”1.0” encoding=”UTF-8” standalone=”no”?>
<SOAP-ENV:Envelope
SOAP-ENV:encodingStyle=”http://schemas.xmlsoap.org/soap/encoding/”
xmlns:SOAP-ENV=”http://schemas.xmlsoap.org/soap/envelope/”>
<SOAP-ENV:Body>
<m:AddIt xmlns:m=”http://tempuri.org/message/”>
<First>1</First>
<Second>2</Second>
<Result>3</Result>
</m:AddIt>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>

Notice that this output is relatively uncluttered and looks very much like the simple exam-
ple we talked about earlier. While this output works with Microsoft products, it doesn’t nec-
essarily provide full compatibility with the SOAP specification. We’ll talk about how this
54 Chapter 2 SOAP In Theory

sample misses the specification and what Microsoft plans to do about it as the section
progresses. The second listing is from the 4S4C (Simons Soap Server Services for COM).
<?xml version=”1.0”?>
<SOAP-ENV:Envelope
xmlns:SOAP-ENV=”http://schemas.xmlsoap.org/soap/envelope/”
xmlns:SOAP-ENC=”http://schemas.xmlsoap.org/soap/encoding/”
xmlns:xsd=”http://www.w3.org/1999/XMLSchema”
xmlns:xsi=”http://www.w3.org/1999/XMLSchema-instance”
SOAP-ENV:encodingStyle=”http://schemas.xmlsoap.org/soap/encoding/”>
<SOAP-ENV:Body>
<m:AddIt xmlns:m=”http://WinServer/4s4c/MyAddIt”>
<defaultname xsi:type=”xsd:int”>3</defaultname>
<First xsi:type=”xsd:int”>1</First>
<Second xsi:type=”xsd:int”>2</Second>
</m:AddIt>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>

While 4S4C produces complex looking output, it also provides more information than the
Microsoft SOAP Toolkit. In many cases, the additional output can mean the difference
between an application that works and one that doesn’t. Of course, the 4S4C toolkit has
problems of its own. The biggest problem is that the XML tag doesn’t contain the encoding
attribute, which confuses the parsing mechanisms for some languages. Since this problem is
relatively minor and only appears in one place at the beginning of the message, you could
theoretically hand code around it for now. Hopefully, the 4S4C author will fix this problem
with the next release of the product.
You need to consider several SOAP specific issues when looking at the output of these two
toolkits. According to the SOAP specification, the placement of the result in the Microsoft
message is incorrect—it should appear as the first element in the message, not the last as
shown. The current ordering is more like the output of COM, than the output of SOAP. It
also means that SOAP message decoding is different for the Microsoft SOAP Toolkit output
than for 4S4C. This ordering issue can cause problems with clients who are expecting to
receive elements in a specific order. Microsoft says they’ll fix this problem and they may have
done so by the time you read this. Still, element ordering is an important issue to consider.

4S4C is just one of many SOAP toolkits on the market that allow you to create SOAP
application using Visual C++. We’ll cover more of these toolkits as the book progresses.
Look at http://www.pocketsoap.com/ 4s4c/ if you want to learn more about
4S4C. 4S4C is an easy to use toolkit that includes many features that the Microsoft
offering doesn’t. For example, this toolkit offers complex variable handling and custom
interface support. It also provides superior support for Visual C++ users—at least as of
this writing.

Another problem with the Microsoft SOAP Toolkit may not go away very soon. Notice that
the Microsoft Toolkit doesn’t provide proper qualification attributes showing the type of
the input values (First and Second) or the return type. This particular problem causes the
 Using SOAP to Move Data 55

output of this toolkit to fail when used with an Apache server. In this case, Apache promises
to fix their server offering since the SOAP specification doesn’t require type attributes for
variables. However, using type attributes is still a good idea since it avoids ambiguity prob-
lems and makes it easier to decode the SOAP message. The price, of course, is increased
SOAP message size, which may affect performance.
The Microsoft SOAP Toolkit also uses “Result” as a return value name, no matter what the
component’s author may have requested. While the SOAP specification isn’t very clear
about this problem, the common way of handling a return value without a name is to use
“defaultname” as shown in the 4S4C example. This particular problem only exacerbates the
problem of element ordering since it’s impossible to specify an element by name if there Ch
isn’t a common naming method. Unfortunately, the SOAP committee will need to work out
this problem before vendors will know how to fix the problem in their toolkits.
2
The envelope tag in the Microsoft SOAP Toolkit example is missing two important attrib-
utes, which include the XML Schema Definition (XSD) location and the XML Schema for
Interfaces (XSI). Parsers need both of these attributes if you want them to define variable
types throughout the SOAP document. While the SOAP specification doesn’t say that a
SOAP implementation must provide these attributes, it does show them throughout the
entire document including the example code. Again, Microsoft has promised to fix this prob-
lem in a future build of its toolkit. The important consideration is that you’ll want to check
for these entries if you’re having a problem with data typing within a SOAP document.

Like everything else with SOAP, you need to define the location of the type attribute
specifications or schemas. You typically define two namespaces to make this work.
The first, xmlns:xsd=”http://www.w3.org/ 1999/XMLSchema”, is the XML
Schema Definition. The second, xmlns:xsi=”http://www.w3.org/1999/
XMLSchema-instance”, is the XML Schema for Instances. You can find a complete
write-up of both entries at http://www.w3.org/TR/xmlschema-0/. This docu-
ment also tells you which data types XML supports natively. Note that your SOAP
implementation may not support all of these data types right now, but probably will in
the future. You’ll also want to learn about the other two parts of the XML schema:
structures at http://www.w3.org/TR/xmlschema-1/ and data types at
http://www.w3.org/TR/xmlschema-2/.

By this time, you may be concerned that you’re going to run into compatibility problems
when working with SOAP. It’s important to remember that SOAP is still in its infancy. The
SOAP committee will eventually develop a robust specification. Vendors will eventually
implement that specification in a way that allows SOAP projects to interact.
Of course, there’s always the specter of vendor bragging rights and marketing with which to
contend. Vendors continue adding extensions to standardized products in an effort to differ-
entiate their product from someone else’s. The main purpose of this section of the chapter
isn’t to present you with a hit list of problems in SOAP implementations today—it’s to point
out problems that you should look for in the future. Creating a well-formed SOAP docu-
ment might seem easy, but, as shown here, is sometimes harder than it looks.
56 Chapter 2 SOAP In Theory

SOAP and XML-RPC

XML-RPC was one of the first text-based remote procedure call (RPC) technologies on the
market designed for Internet use. If you spend some time looking at this protocol, you’ll
notice that it bears some resemblance to SOAP, but the resemblance isn’t strong. XML-
RPC relies on HTML type organization within an XML framework.

There are several places on the Web to find out more about XML-RPC. The XML-RPC
home page is at http://www.xml-rpc.com/. You can also find a copy of the cur-
rent specification at http://www.xmlrpc.com/stories/storyReader$7. A
good overview and simple explanation of XML-RPC appears at http://davenet.
userland.com/1998/07/14/xmlRpcForNewbies/. You’ll also find an interesting
XML technology comparison chart at http://www.w3.org/2000/03/29-XML-
protocol-matrix.

One of the big differences between XML-RPC and SOAP is the lack of namespaces within
XML-RPC. Everything needed to define a component and its interface appears within the
document. This organization is simpler than SOAP at the outset, but the lack of flexibility
can create problems. If XML-RPC doesn’t know how to handle a new variable type, you
can’t add support for that variable type. SOAP is far more extensible than XML-RPC.
Interestingly enough, XML-RPC does provide support for complex data type like structures
and arrays. It’s also interesting to note that XML-RPC supports an encoded binary data
type. If you want to get binary data type support with SOAP, you have to use an attachment.
XML-RPC relies on HTTP for transport purposes like SOAP. It also uses a similar, though
simplistic, fault mechanism. Again, the extensibility of SOAP provides you with more error
reporting options.
You can convert XLM-RPC messages to SOAP messages. Since both are text messages,
there aren’t any binary compatibility problems to consider. I’m not going to provide you
with a blow-by-blow description of the process in this chapter, but you can find a write-up
on the topic at http://lwprotocols.org/xmlrpc2soap.html.

SOAP and XML Protocol (XMLP)

This is the new kid on the block. At the time of writing, the vendors supporting this proto-
col had barely begun their work. I wanted to include this protocol because it looks like it will
become an important tool for most developer toolboxes. Of course, you’ll want to wait until
the XMLP protocol becomes a standard before you spend much time developing for it.

You can find out more about XMLP at http://www.w3.org/2000/xp/. Another

interesting place to find general information about XMLP is http://www.oasis-
open.org/cover/xp-prot.html.
 Understanding SOAP Attachments 57

The element of XMLP that I find interesting is that it’s a peer protocol technology. Peer-to-
peer communication has gained interest again because of the way companies like Napster
are using it.
XMLP is also one of the few RPCs to provide more than one transport protocol. You can
use XMLP on both HTTP and SMTP. Even one additional protocol tends to make a proto-
col more flexible. Of course, most people will probably use XMLP on HTTP for now.
You may not use XMLP for your next project—it’s still a new protocol in the midst of cre-
ation. However, XMLP shows a great deal of promise and you should keep your eye on it
for the future. The model on which XMLP is based should provide maximum flexibility and
Ch
extensibility (two catch phrases for any new technology today).
2
Understanding SOAP Attachments
Some people tend to forget that HTTP is a transport mechanism only—HTTP doesn’t
define the content of the message that it carries. That’s why we can use HTTP to transport
SOAP messages from one place to another. Likewise, SOAP, which vendors have based on
XML, only defines one potential type of message within the HTTP envelope. Just as bill
collectors commonly place advertising in your billing statement, vendors can place other
materials within the HTTP envelope that carries a SOAP message. In fact, that’s precisely
what this section of the chapter talks about.
The secret to SOAP attachments is using the Multipurpose Internet Mail Extensions
(MIME). This is the same magic used to send attachments in mail. Anyone who has
browsed their e-mail in source code mode knows how a MIME message is put together.

Unlike many of the extensions to a specification that you’ll run into, there’s a specifica-
tion for SOAP attachments. You can read about it at http://msdn.microsoft.com
/xml/general/soapattachspec.asp. At the time of writing, this particular part of
SOAP isn’t part of a public specification process. However, given the way Microsoft and
other vendors are handling SOAP, there’s little reason to doubt that SOAP attachments
will become a public specification and approved by a standards group. In addition, this
technique relies on existing technology, so there’s little reason to think that a standards
group won’t approve the specification.

Let’s dissect a MIME message for the sake of discussion. A mime message consists of at
least three entries.
1. A MIME header
2. A MIME part description
3. Content

The three pieces will always begin in this order. Of course, the reason to use MIME is
that you can create a message with multiple parts. Parts 2 and 3 of the message can repeat
58 Chapter 2 SOAP In Theory

several times—Part 1 only appears once at the beginning of the message. Here’s an example
of Part 1 of the message.
MIME-Version: 1.0
Content-Type: Multipart/Related; boundary=MIME_boundary; type=text/xml;
start=”<[email protected]>”
Content-Description: This is a sample message header.

The Content-Type entry in the source code above normally appears on one line. I’ve
shown it on two lines to accommodate the formatting requirements of the book.

Note that this part description contains the MIME version number, the type of content, a
starting location, and a description of the content. The starting point is especially important
because it’s the first thing the recipient will see. The “start=” attribute always points to the
SOAP message. Here’s an example of Part 2 of the message.
--MIME_boundary
Content-Type: text/xml; charset=UTF-8
Content-Transfer-Encoding: 8bit
Content-ID: <[email protected]>

As you can see, this part uses the keyword “MIME boundary” to show that it’s another part
of the message. There are also content type, charset, and encoding entries, much like the
HTTP header we looked at earlier in the chapter. The Content-ID entry defines the name
of this part of the message. The first part of the Content-ID entry defines the name of the
file, while the second part contains the domain that generated the content.
The third part of a MIME message sequence is the data. We’ve already looked at SOAP
messages several times in the chapter, so I won’t repeat that information here. There isn’t
any difference in the SOAP format for a MIME message—it’s the same as a standalone mes-
sage. Of course, you can just as easily include any binary file as part of the MIME message.
Most people will associate MIME with email messages. However, MIME, like SOAP, can rely
on any of several transport protocols including HTTP. The Microsoft documentation shows
you how to bind MIME with HTTP. In this case, you replace Part 1 of the MIME message
sequence with a special type of HTTP header. The header looks similar to the other HTTP
headers discussed in the chapter so far. The main difference is the Content-Type entry looks
like the one used for a MIME message, rather than the typical HTTP header entry.

Case Study
Sometimes a SOAP solution takes the form of cooperative business. Take, for example, one
company that we’ll call MyBuy Corporation that decided to set up a merge-in-transit system.
A merge-in-transit system is where two companies cooperate to fulfill customer orders in
the most efficient manner possible. One company produces goods, while another company
assembles goods from various factory outlets for shipment to customers. You’ll see how this
works as the case study progresses.
 Case Study 59

In this case, MyBuy and MyMail (a fictitious name for a company like FedEx or UPS)
decided to team up. MyBuy receives orders from customers for a variety of their products.
Since MyBuy is nationwide, they need warehouses in several states to provide timely delivery
of their products. The problem is that these warehouses are costly to operate and maintain.
Not only that, but MyBuy also pays MyMail to pick up orders at the closest warehouse for
delivery to customers—another expense. Figure 2.4 shows the old company setup—you can
see how inefficient it is.

Figure 2.4
The original MyMail MyBuy Factories
setup reflects what Ch
most companies
use today, which is 2
inefficient and costly.

MyBuy Factory
Warehouses

MyBuy Assembly
Warehouse

MyBuy Local
Warehouses

Customers

Someone at MyBuy decided that the company could eliminate 3% to 5% of its business
expenditures by getting rid of all of those local warehouses. Using a factory direct approach
would mean that MyBuy wouldn’t need to keep track of inventory in each warehouse in
hope of fulfilling orders from existing stock.
This is where MyMail comes into play. MyMail has to store MyBuy’s products in their ware-
houses anyway before shipment. In addition, MyMail has to maintain local warehouses to
provide service to all of their customers. When MyBuy approached MyMail about their
problem, the two companies decided to work together to solve it. MyBuy eliminated their
local warehouses and began to store their products at the extensive MyMail local warehouses.
MyBuy simply sent products from their factories to the MyMail assembly warehouses where
60 Chapter 2 SOAP In Theory

MyMail employees put orders together and ship them to customers. Figure 2.5 shows what
this arrangement looks like.

Figure 2.5
The new MyMail MyBuy Factories
setup saves money by
reducing the number
of distribution points.

MyMail Assembly
Warehouse SOAP Orders

MyMail Local
Warehouses

Customers

The only problem with this arrangement, of course, is that the two companies need some
way of passing order information to each other. That’s where SOAP comes into play. When
MyBuy receives a customer order, an employee keys the information into a local database.
The order information travels from the MyBuy company site to the MyMail Web site using
the same business-to-business transaction Web site that other MyMail customers use. The
transaction wouldn’t be possible without SOAP because the two companies use different
computer setups and their databases don’t accept data in the same way. A translation has to
take place—one that isn’t easy to perform using older technologies like DCOM. (Even if
MyBuy and MyMail could agree on a single setup, MyMail has other customers to support.)
MyMail actually gets the order shipped a day earlier than normal because the MyBuy ware-
house doesn’t have to assemble the order and then call MyMail to pick it up. As a result,
the customer normally receives their order earlier. Not only is MyBuy more efficient, but
customer satisfaction increases as well.
However, there’s an added bonus in this setup. When a customer uses the MyBuy Web site,
the order goes directly from MyBuy to MyMail without any intervention. This means that
an order made by a customer at 1 a.m. could theoretically arrive the same day. The MyMail
warehouse that’s closest to that customer will receive the order and could fulfill it if all of
the needed parts from the MyBuy factories are on hand.
 CHAPTER
3
An Overview of Security Issues
for SOAP
In this chapter
Introduction 62
Understanding SOAP Privacy and Security Issues 63
Security Standards You Should Know About 67
User Identification Issues 75
Where Do You Go from Here? 77
Case Study 78
62 Chapter 3 An Overview of Security Issues for Soap

Introduction
Security figures prominently in the headlines today. It seems that security is always an issue
for someone. The story is always the same. A cracker breaks in, causes some problems, an
administrator in the target company catches them, and the company promises to do some-
thing about security in the future. Of course, fixing the problem after someone breaks into
your system is like putting your seat belt on after an accident. Yes, you can prevent future
damage, but your system has already suffered loss.

For the purposes of this book, the term cracker always refers to an individual that’s
breaking into a system on an unauthorized basis. This includes any form of illegal
activity on the system. On the other hand, a hacker refers to someone who performs
low-level system activities, including testing system security. In some cases, you need to
employ the services of a good hacker to test the security measures you have in place,
or suffer the consequences of a break-in. This book uses the term hacker to refer to
someone who performs these legal forms of service.

This chapter is going to provide a basic overview of common security considerations for
SOAP applications. We’ll talk about security quite a bit as the book progresses and we get
into the example programs. Consider this chapter as the introduction to more material. It
talks about common elements, places to get additional help, and some of the tasks you need
to perform for each application.
Every developer should remember two things as part of the development process. First,
security is an ongoing process. As soon as you fix one set of problems, crackers will find yet
another way to break into your system. Therefore, it’s important to make your applications
flexible. Make sure you can add new security features as needed. Second, no matter how
well you designed your application, someone will figure out a way to break the security
measures you included with it. With this in mind, you should consider adding a level of
security monitoring with the application. We’ll talk about this particular issue as the book
progresses because security monitoring is a customized part of the programming process.
The first section of this chapter, “Understanding SOAP Privacy and Security Issues” dis-
cusses how working with a public network is different from working with applications on
the LAN. One important issue is privacy. Not only do you need to keep company data safe
from prying eyes, but you also need to provide customer identities and other information
that usually aren’t exposed on a LAN. Given the text-based nature of SOAP transfers, this
issue becomes even more important.

Microsoft and other vendors realize that security is an increasing concern because of
the number of cracker attacks hitting both their Web sites and those of their customers.
Some vendors now talk about levels of security service and the need to differentiate
these levels by vendor. For example, Microsoft knows that its current security policies
 Understanding SOAP Privacy and Security Issues 63

leave something to be desired in a world of distributed applications, so they’re imple-

menting new security strategies to overcome some of these limitations. You can read
about Microsoft’s new third-wave security strategy at http://www.microsoft.com
/technet/security/ thrdwave.asp.

The second section, “Security Standards You Should Know About,” discusses the important
standards in use today. Sometimes, a developer will limit the sources of information used to
make security decisions. Using standards-based security ensures you get the best security
available that will work with solutions used by other companies. In addition, knowing the
standards is like placing additional tools in your toolbox. You need this information to do a
great job putting your next application together.
The HTTP Authentication Framework, Secure/Multipurpose Internet Mail Extensions
(S/MIME), and Secure Socket Layer (SSL) sections explore ways to keep your SOAP appli-
cation secure. Each of these technologies will add to your ability to keep both company and
Ch
user data safe. You’ll need to use a security technology that matches your data transfer strat-
egy. For example, S/MIME works best when you need to use SOAP attachments. (We dis- 3
cussed SOAP Attachments in the “Understanding SOAP Attachments” section of Chapter 2.)
The “User Identification Issues” section talks about methods for determining user identify.
You don’t want just anyone to access your Web site. Unfortunately, given user privacy issues,
you might have a hard time determining that a user is who he or she says they are.
Fortunately, there are some ways around this issue, especially if security is a big concern.
The “Where do You Go from Here?” section provides some pointers on what you can do to
improve the security of your SOAP application in general. Remember that security is an on-
going process and your application needs to change to keep up with security requirements.
The final section of the chapter is a security-related case study. We’ll look at the theory
we’ve talked about in this chapter in action at a company just like yours. I’ll include some
tips on how this company could have avoided some of the problems that it encountered.
This section also discusses what you can do when preventative care won’t do the job and
you need to resort to monitoring to catch crackers as they enter your network.

Understanding SOAP Privacy and Security Issues

Anyone making even the smallest survey of trade press and magazine articles today will
notice a definite increase in security and privacy articles. Corporations are tightening their
security perimeters by purchasing an increasing number of security products. Security is a
big money issue today; there isn’t any doubt about it. In addition, privacy issues are making
their rounds through government agencies and it isn’t clear how the government will
respond to growing pressure for individual privacy protection.
Of course, security vendor hype has something to do with increased panic in corporations.
I used to get an occasional e-mail detailing some new virus. Today I get a new virus notice
64 Chapter 3 An Overview of Security Issues for Soap

almost every day. Eventually I may have to filter my messages to stem the tide of messages
filled with woe and gloom. I take viruses seriously and use preventative software, but some-
times the hype factor verges on the ridiculous for viruses that haven’t even become a threat.
Frontpage stories of identity stealing and other privacy issues don’t help. People are concerned
that someone will steal their identity with the very next purchase they make online. The gov-
ernment is weighing the right someone has to privacy against their desire to purchase items
online. In the meantime, a few vendors are playing on people’s fear by selling them privacy pro-
tection policies. If someone steals their identity, the company promises to help them recover it.
Certainly, hype is a problem in the computer industry. If we weren’t suffering through security
hype attacks, then vendors hawking other wares would hit us with some type of hype. Hype
aside, security and privacy are serious issues. The following sections address important security
and privacy issues in a way that helps you cut through the hype and concentrate on getting
that next application completed.

What Are the Issues?

One of your first tasks in creating a distributed application is figuring out where hype ends
and reality begins. On one hand, you can’t afford to ignore security and privacy issues
because governments are creating laws that say you must care or spend time in jail. On the
other hand, you don’t want to spend all your time in a dither, wringing your hands and won-
dering when the government will break down your door. Distributed computing opens a
new world of security problems, but you don’t have to be overwhelmed by them. The fol-
lowing list includes reasons you need to care about security beyond desktop application
requirements when building distributed applications.
■ Crackers are becoming more persistent. The wealth of virus kits just makes their job
easier. Your data is definitely a prime target for corruption by cracker attack. For that
matter, they might just decide to steal your data and sell it to the highest bidder.
■ Using a text-based data format makes your application work with other vendor’s systems
better, even if those systems don’t rely on the same hardware or operating system. Of
course, crackers can easily read and modify text data without leaving any fingerprints.
Good security relies on placing roadblocks against intrusion and then detecting times
when the roadblocks don’t work. Text-based protocols make detection impossible and
protection improbable.
■ Governments all over the world are creating laws that address both corporate and per-
sonal privacy. You’re legally responsible for protecting user data from prying eyes—a
difficult task even in the best situation.
■ Leaving ports open on your system allows partners, customers, and employees on the
road access to your system. These open ports are also an invitation to cracker attack.

I could make this list much longer. The point is that distributed applications face many
threats that desktop applications will never face. When you work with a desktop application,
there’s a certain level of physical security involved that you don’t receive when working
online. In addition, the threats you see today will become old and mundane while new threats
 Understanding SOAP Privacy and Security Issues 65

take their place. Security constantly changes when working with distributed applications
because new vulnerabilities are exposed and as you learn new techniques to provide security,
crackers are learning new ways to break through. You need a three-part plan to protect your
data that includes security monitoring, improved security measures, and training.
To give you an idea of just how much the security scenery changes, consider the problems
with the Berkeley Internet Name Domain (BIND) protocol. This protocol allows Domain
Name Service (DNS) servers to translate human readable Web site names to IP addresses.
Who would have thought that crackers would use a hole in this protocol to bring
Microsoft’s servers down. The crackers weren’t, in fact, aware of the problem until a
Microsoft technician made a configuration mistake that took Microsoft’s Web site offline.
Now that crackers know about the security vulnerability, Web site operations will need to
upgrade to a clean version of BIND or suffer the consequences.

User Privacy Issues

User privacy is going to become a major headache for most developers because there are so
Ch
many government factions getting into the picture. The one thing that you should count on
doing is making any user data exchanges airtight. You need to ensure databases are secure 3
and build a buffer between them and the Internet, if possible. Data aging is another impor-
tant issue. Any application you build should provide some sort of user-information aging so
it eliminates users who don’t visit on a regular basis.
However, no matter how many precautions you take, some cracker is going to break in if he or
she is determined to do so. I read articles in the trade press about application failures and infor-
mation releases due to unknown security holes on a regular basis and I imagine that you do too.
It’s unlikely that the people in the article could have foreseen the security breach or done any-
thing more to protect against it. In fact, these articles seldom lambaste the people in question
for the security problem or the failure of their software. What the article usually points
out is that the company denied the problem and failed to notify users of the security breach.
Modern applications should include some form of security monitoring. I’m not talking about the
big brother type of monitoring, but the type that alerts a network administrator to irregularities
in data processing. Something as simple as reporting a user who continually tries multiple times
to access the Web site, only to fail, can help locate crackers trying to break into your system.
The application should maintain logs and provide some form of statistical analysis so that the
network administrator can look for trends. Of course, you have to do all of this within the guide-
lines of the new laws that governments pass to protect the identity of users who access your site.

At least two vendors are working on new standards that will help protect user privacy,
as well as the content of a SOAP message. The first is the Security Service Markup
Language (S2ML). You can find a good general page for this standard at http://xml.
coverpages.org/s2ml.html. The second is AuthXML. You can find a good general
page for this standard at http://xml.coverpages.org/authxml.html. In both
cases, the standards allow a vendor to add attributes to an XML (including SOAP) mes-
sage that describe access rights for each element. These standards also provide for
66 Chapter 3 An Overview of Security Issues for Soap

authentication and other important privacy issues. (Note that the standards groups for
both of these standards appear in Table 1.1.)

User privacy often treads on the verge of the bizarre for the developer. One of the privacy
issues that developers and the legal minds of government need to consider is how to treat
user information when used in an official capacity. For example, if you’re an individual
shopping at Amazon.com, then the rules for sharing your information are pretty clear—the
vendor has to keep your information safe unless you grant permission to give it away.
However, what happens when a corporate office makes a business purchase over the Internet?
In other words, it’s really a company making the purchase; the individual is simply the inter-
mediary between the two companies. In addition, the individual supplies their company
identity, not his or her personal one. Does the law protect a person’s corporate identity in the
same way as his or her private identity? The answer is unclear right now, but it’s an important
question to consider. As a developer, you may find yourself writing special code for those busi-
ness-to-business transactions you coded using standard client/server techniques in the past.
Another privacy issue is the matter of opting into a mailing list versus opting out of it. Many
vendors would like to use the opting out scheme, which means they could assume everyone
wants to be on their mailing list unless they specifically say otherwise. Vendors state that
they’ll maintain larger lists using the opt out option because fewer people are likely to ask to
leave the list than take the time to opt into a mailing list. In other words, the vendors will
be able to sell their list to other companies at a higher price because it will contain more
names. Privacy groups, on the other hand, want to adopt the opt in approach because it
offers greater individual protection. Again, this is an important issue to consider because
it affects how you write your application

Data Security Issues

You’ve likely read countless tomes on maintaining the security of applications in general.
For example, I probably don’t need to tell you that you should password protect your appli-
cation or that you should only assign the level of security a user actually requires. Windows
2000 even adds the new role-based security model in which you can assign security to indi-
vidual components and even methods within an application. All of these measures work well
within an internal application because you don’t have to worry about outside influences
nearly as much as you do about disgruntled employees.
Distributed applications require additional levels of security that internal applications don’t
require. For example, anyone on the Internet can intercept the data flowing between your
site and the site of your customer. You need to encrypt the data in some way to prevent prying
eyes from looking at it. Unfortunately, the technologies required to encrypt the data also
cause problems with firewalls. DCOM and CORBA both provide encrypting technology
that’s hard to break because they encrypt the entire message (sometimes at several levels).
 Security Standards You Should Know About 67

Unfortunately, neither of these protocols will do you much good because they don’t work
very well on the Internet. We’ll see later in the chapter that some companies are working on
ways to encrypt SOAP messages, but this technology is in its infancy as I write this.
Now, consider another problem. You encrypt the data using a well-known scheme.
Unfortunately, the other party using your application doesn’t have the same platform. As a
result, they don’t have access to your well-known encryption scheme and can’t decrypt your
message. Some vendors are working on a common encryption scheme, but these common
schemes usually end up working on a limited number of platforms. As a developer, you nor-
mally need to provide access to several different encryption schemes, so the endpoints of
your application can decide on an encryption scheme that will work for everyone.
Disparate operating systems and hardware also present another problem—data translation. I’m
unaware of any method for translating data from one format to another while in an encrypted
state. Decrypting the data to translate it may present a problem if done incorrectly. You’ll want
to ensure your application performs any required translations in a safe place, which means
behind a firewall and on a secure server. Of course, the very act of translating the data opens Ch
the possibility of introducing data errors. SOAP partially mitigates this problem because it
uses a text format that servers can easily parse. 3
A final problem to consider is one of technological differences between platforms. Consider
the current problem with wireless technology. Vendors admit that the current technology
leaves data unprotected during the transition from the world of wireless communication to
the LAN. These vendors are working on a solution, but again, at the time of this writing
none exists. The time the data remains unprotected is short, but crackers have taken advan-
tage of less serious openings in the past.

Security Standards You Should Know About

Security standards are an important part of your company’s safety net. Most administrators
know there are holes in the security net for their company. These holes come from a variety
of sources, including the operating system, ports open to the Internet, and off-the-shelf
applications that they rely on to conduct business. All of those problems are real, but you
don’t have to face them alone. Standards groups are working even as you read this to come
up with methods for protecting data. All you need to do is learn the methods that these
groups come up with for managing security on your network.
The advantages to using standards-based security are twofold. First, you won’t have to
reinvent the wheel and create everything from scratch. Second, your security methods will
mesh with those used by other sites, reducing the user learning curve and making it possi-
ble for you to use tools developed for other programmers. (There’s a rather dubious third
advantage that you’ll have some idea of where security breaches might appear and can build
components to monitor for them until the associated standards committee defines a fix for
the problem.)
Let’s look at some of the existing and proposed security standards that affect the Internet
today. Table 3.1 contains a list of what I consider essential security standards. Notice that
68 Chapter 3 An Overview of Security Issues for Soap

the Internet Engineering Task Force (IETF), http://www.ietf.org/, and World Wide Web
Consortium (W3C), http://www.w3.org/, play a major role in devising and implementing
Internet standards (with the help of member vendors, of course). You can find most of the
IETF RFC documents at http://www.rfc-editor.org/. You can also find lists of the current
IETF working groups at http://www.ietf.cnri.reston.va.us/html.charters/. These
working groups help create the standards used on the Internet. The W3C usually has a list
of their standards on their main Web page, so you can just click a single link and get to
where you need to go.

Table 3.1 Internet Security Standards and Specifications

Standard Description
AuthXML This security technology is in the specification stage as I write this—
the vendor does intend to submit it to a standards committee (W3C
most likely). AuthXML is an authentication and authorization speci-
fication that would replace localized forms of these same technologies.
It’s based on XML and relies on digital signatures to perform its
work. Since this isn’t an encoding specifica tion, any data you trans-
mit is still open for interception by third parties. You can find out
more about AuthXML at http://www.authxml.org/.
Distributed Authentication DASS defines an experimental method for providing authentication
Security Service (DASS) services on the IETF RFC1507 Internet. The goal of authentication,
in this case, is to verify who sent a message or request. Current pass-
word schemes have a number of problems that DASS tries to solve.
For example, there’s no way to verify that the sender of a password
isn’t impersonating someone else. DASS provides authentication ser-
vices in a distributed envi ronment. Distributed environments pre-
sent special challenges because users don’t log onto just one
machine—they could conceivably log onto every machine on the
network. You can find out more about this standard at http://www.
wu-wien.ac.at:8082/rfc/rfc1507.hyx/$$root.

Digital Signatures This is a standard originated by Initiative (DSI) W3C to overcome

some limitations of channel-level security. For example, channel-level
security can’t deal with documents and application semantics. A
channel also doesn’t use the Internet’s bandwidth very efficiently
because all the processing takes place on the Internet rather than at
the client or server. DSI defines a mathematical method for transfer-
ring signatures— essentially a unique representation of a specific
individual or company. Find out more at http://www.w3.org/
DSig/Overview.html. It’s interesting to note that there’s an XML
version of this technology underway. Read about this new technology
at http://www.w3.org/Signature/.
eXtensible Rights Markup A new ContentGuard specification that defines how someone can use
Language (XRML) content provided by your company. It describes the rights, fees, and
conditions of use for content. XRML also allows a vendor to define
trusted systems that can use a product for testing and evaluation pur-
poses. This technology relies on a trusted server to determine if some-
one can access content and what rights they have when they do. You can
learn more about this specification at http://www.xrml.org/.
 Security Standards You Should Know About 69

Table 3.1 Continued

Standard Description
Generic Security Service This specification defines methods for supporting security service
Application Program calls in a generic manner. Using a generic interface allows greater
Interface (GSS-API) source code portability on a wider range of platforms. IETF doesn’t
IETF RFC1508 see this specification as the end of the process, but rather the starting
point for other, more specific, standards in the future. However,
knowing that this standard exists can help you find the thread of com-
monality between various security implementation methods. You can
find out more about this standard at http://www.wu-wien.ac.
at:8082/rfc/rfc1508.hyx/$$root.

HTTP Authentication One of the problems with HTTP 1.0 authentication is that endpoints
Framework RFC2617 send passwords in the clear, making it easy for others to grab the
passwords and circumvent security. The HTTP Authentication
Framework provides a means for encrypting authentication informa-
tion—improving overall security. You can find out more about this
standard at http://www.faqs.org/rfcs/rfc2617.html.
Ch
Internet Protocol This specification addresses issues of IP client security, such as the
Security Protocol (IPSec) inability to encrypt data at the protocol level. It includes a wide 3
range of specifications that will ultimately result in more secure IP
transactions. You can find out more at http://www.ietf.cnri.
reston.va.us/html.charters/ipsec-charter.html.

Private Communication Like SSL, the IETF designed PCT to provide a secure method
Technology (PCT) of communication between a client and server at the low-protocol
level. It can work with any high-level protocol such as HTTP, FTP,
or TELNET. You can find out more at http://www.graphcomp.
com/info/specs/ms/pct.htm.

Privacy Enhanced Mail This specification outlines a procedure for encrypting mail in a way
Part I (PEM1) Message that protects the user’s mail but the process of decrypting it is
Encryption and invisible. This includes the use of keys and other forms of certificate
Authentication Procedures management. Some of the specification is based on the CCITT
IETF RFC1421 X.400 specification—especially in the areas of Mail Handling Service
(MHS) and Mail Transfer System (MTS). Related standards include:
Privacy Enhanced Mail Part II (PEM2) Certificate-Based Key
Management (IETF RFC1422), Privacy Enhanced Mail Part III
(PEM3) Algorithms, Modes, and Identifiers (IETF RFC1423), and
Privacy Enhanced Mail Part IV (PEM4) Key Certification and
Related Services (IETF RFC1424). Find out more about this stan-
dard at http://www.cs.ucl.ac.uk/research/ice-tel/
osisec/documentation/ or http://www.si.hhs.nl/~henks/
comp/crypt.html (which includes the associated standards).

Secure Multipurpose This specification defines a method for different developers to create
Internet Mail Extensions message transfer agents (MTAs) that use compatible encryption
(S/MIME) technology. Essentially, this means that if someone sends you a mes-
sage using a Lotus product, you can read it with your Banyan prod-
uct. S/MIME is based on the popular Internet MIME standard
(RFC1521). You can find out about standard MIME at http://www.
oac.uci.edu/indiv/ehood/MIME/. There’s a whole list of S/MIME
specific resources at http://www.rsasecurity.com/standards/
smime/resources.html. The S/MIME working group resides at
http://www.ietf.org/html.charters/smime-charter.html.
70 Chapter 3 An Overview of Security Issues for Soap

Table 3.1 Continued

Standard Description
Secure/Wide Area Network The main goal of S/WAN is to allow (S/WAN) companies to mix-and-
match the best firewall and TCP/IP stack products to build Internet-
based virtual private networks (VPNs). Current solutions usually lock
the user into a single source for both products. S/WAN is no longer a
specification or even a work in progress, but it has spawned other
efforts along the same line. Find out more about S/WAN and some of
the efforts associated with it at http://www.rsasecurity.com/
rsalabs/faq/5-1-3.html.

Security Services Markup This is a new data security standard under consideration Language
(S2ML) by Organization for the Advancement of Structured Information
Standards (OASIS). It’s based on XML and designed to provide a com-
mon security language for all programming platforms. The main con-
sideration for this standard is that it’s designed to move RPC packets
such as SOAP messages across the Internet in a secure manner. You
can find out more about this pro posed security standard at http://
www.s2ml.org/.

Secure Hypertext Transfer This is the current encrypted data transfer technology used by Open
Protocol (SHTTP) RFC2660 Marketplace Server, which is similar in functionality to SSL. The big
difference is that this method only works with HTTP. The IETF
formed the Web Transaction Security (WTS) group to look at specifi-
cations like this one. You can find out more about this standard at
http://www.ietf.org/html.charters/wts-charter.html.

Secure Sockets Layer (SSL) This is a W3C standard for transferring encrypted information from
the client to the server at the protocol layer. Sockets allow low-level
encryption of transactions in higher-level protocols such as HTTP,
NNTP, and FTP. The standard also specifies methods for server and
client authentication (although client-site authentication is optional).
SSL is a popular specification that you can find in many places includ-
ing http://webopedia.internet.com/TERM/S/SSL.html (check
the links at the bottom of the page) and http://www.
ncsa.uiuc.edu/InformationServers/WebSecurity/iw3_tut/NETS
CAP1.HTM.

The Kerberos Network The Kerberos model is based in part on Needham and Schroeder’s
Authentication Service (V5) trusted third-party authentication protocol and on modifications
IETF RFC1510 suggested by Denning and Sacco. As with many Internet authentication
protocols, Kerberos works as a trusted third-party authentication service.
It uses conventional cryptography that relies on a combination of shared
public key and private key. Kerberos emphasizes client authentication
with optional server authentication. Find out more at http://info.
internet.isi.edu/in-notes/rfc/files/rfc1510.txt.

Universal Resource A URI provides a means of encoding the names and addresses of
Identifiers (URI) in WWW Internet objects. Currently, resource names and addresses appear
IETF RFC2396 (and others in clear text. An URL (uniform resource locator) is actually a
like RFC1630) form of URI containing an address that maps to a specific location on
the Internet. To visit a private URI site, you’d need to know the
encoded name instead of the clear text name. Technologies like SOAP
use URIs instead of URLs to keep Web site resources private. Learn
more about URIs and how they compare to URLs at
http://www.w3.org/Addressing/.
 Security Standards You Should Know About 71

No, you can’t use all of these standards with SOAP, but you can often supplement your
SOAP coverage by relying on these standards to protect other areas of your network. Two
of the more interesting recent additions to my list of standards are Netegrity’s Security
Services Markup Language (S2ML) and Securant’s AuthXML. Neither vendors (or associ-
ated standards groups) has released these standards, but they’ll appear on the market in the
near future and may provide better protection for SOAP messages than existing technologies
such as SSL. The important thing to remember is that Internet text-based remote procedure
call (RPC) mechanisms provide nothing in the way of native security support, so you need
to use some form of protection for your application. We’ll discuss three of the most impor-
tant (and released) SOAP application security solutions later in the chapter.

If you want to find out the latest information on where the world is going with security
standards (particularly the Internet), check http://www.w3.org/Security/. This page
of general information won’t provide everything you need, but it’ll give you places to look
and links to other sites that do provide additional material. Developers will want to get the
commercial view of security at http://www.rsasecurity.com/. The RSA site covers a Ch
broad range of topics. This site is the best place to start if you want to add new physical
security technologies to your applications like smart cards. You can also find out the status
3
of IETF efforts by viewing the document at ftp://ftp.isi.edu/internet-drafts/
1id-abstracts.txt. Finally, you can find good general information about most XML
technologies, including S2ML and AuthXML, at http://xml.coverpages.org/. It also
pays to look at http://afs.wu-wien.ac.at/usr/edvz/gonter/rfc-list.html
because it organizes standards into functional areas. Make sure you check out the Internet
RFC/STD/FYI/BCP Archives at http://www.faqs.org/rfcs/ if you ever need to find a
request for comment (RFC) document quickly.

HTTP Authentication Framework

Authentication, knowing whom you’re dealing with at the other end of the wire, is one of
the main ways to ensure your communication is safe. Even in the early days of the Internet,
a server could send a 401 (Unauthorized) response message to a client or a 407 (Proxy
Authentication Required) response message to a proxy server in order to start the authenti-
cation cycle. The only problem with these early efforts is that the authentication informa-
tion appeared in clear text. Someone could intercept the password information, keep a copy
for later use, and modify it so the authentication failed. The HTTP Authentication
Framework is associated with HTTP 1.1, both of which provide better means for securing
authentication information between a client and server. One of the major reasons to use the
HTTP Authentication Framework is to ensure both parties encrypt passwords and other
user authentication information.
So, how does the HTTP Authentication Framework help a SOAP developer? First, it tells
you that you should use HTTP 1.1 whenever possible to ensure private communications.
While HTTP doesn’t provide the best protection for your application, it does provide some
72 Chapter 3 An Overview of Security Issues for Soap

level of protection. As a minimum, you can protect all important user authentication infor-
mation. Of course, the HTTP Authentication Framework isn’t a data encryption standard.
Your data is still open to prying eyes—only the authentication portion of the communica-
tion is secure.
The mechanism behind the HTTP Authentication Framework couldn’t be simpler. A client
requests a resource such as a Web page from the server. The server sends a 401 or 407
response in place of the 200 (OK) response that it normally sends. The response also
includes a challenge entry in the HTTP header that indicates the type of challenge the
server poses and credentials entry that tells what type of credentials the client requires.
Here’s a simple example of the entries you’ll see in the response header.
HTTP/1.1 401 Unauthorized
WWW-Authenticate: Digest
realm=”[email protected]”,
qop=”auth,auth-int”,
nonce=”dcd98b7102dd2f0e8b11d0f600bfb0c093”,
opaque=”5ccc069c403ebaf9f0171e9517f40e41”

Of course, there may be other entries in the header—these entries represent what you’ll see
for the HTTP Authentication Framework. The response header begins with the HTTP
version number and status, as mentioned earlier. The WWW-Authenticate entry indicates
that the server is responding with a digest of information. This digest includes the URI of the
realm (resource) in question. The quality of protection (QOP) entry comes next. It includes
two tokens in most cases: auth indicates that this is an authentication header and auth-int
indicates that this is an authentication integrity check. The nonce entry is a unique string that
the server creates for each message using a special algorithm. The client uses this string to
ensure that the response message is unique—not one repeated by a cracker. Finally, the
opaque entry is another string generated by the server. In this case, the client uses the string
within its response to the server. In short, this is the client’s encrypted identification for that
session so the server knows it’s not receiving input from another party (such as a cracker).
The client displays a password dialog box for the user. After the user enters a username and
password, the client sends this information in encoded format to the server. The client sends
the information as token and value pairs. Note that the authentication information appears
as an encrypted string. Here’s a simple example of what you might see.
Authorization: Digest username=”Mueller”,
realm=” [email protected] “,
nonce=”dcd98b7102dd2f0e8b11d0f600bfb0c093”,
uri=”/mydir/default.htm”,
qop=auth,
nc=00000001,
cnonce=”0a4f113b”,
response=”6629fae49393a05397450978507c4ef1”,
opaque=”5ccc069c403ebaf9f0171e9517f40e41”

The client request message contains many of the same elements as the server response.
Notice that the realm, nonce, and opaque values are the same. The QOP has the same
meaning, but a client can’t use the auth-int token. The client must provide a username as
 Security Standards You Should Know About 73

part of the digest that uniquely identifies the user on the server (not the client machine).
The uri entry contains the name and path of the resource the client wants to access. The
reason for this entry is that proxies can act as an intermediary for the client and this entry
reflects the client wishes, rather than those of the proxy. The NC entry contains a nonce
count, which reflects the number of times that a client has received the current nonce from
the server. The cnonce contains a unique string generated by the client. The server uses this
string to authenticate the client (not the user). It also helps in detecting plaintext attacks by
third parties. The client only provides a cnonce value if the server provides a QOP entry.
The response value contains the username and password information the server requested in
encrypted format.
The server validates the client against a security database. If the user information is correct,
the server sends the requested resource and includes the 200-status value. (See Figure 2.1
and accompanying description for a detailed look at how HTTP encases a SOAP message.)
The HTTP Authentication Framework includes the concept of realms. A developer can
partition a resource and identify each partition with a different realm. The user requesting Ch
the resource will receive only the realms that he or she is qualified to receive.
3
Secure/Multipurpose Internet Mail Extensions (S/MIME)
In the “Understanding SOAP Attachments” section of Chapter 2, we discussed how you
could use MIME to create multipart SOAP messages that contain attachments. Using
MIME allows you to send attachments using SOAP—even binary ones. However, it does
nothing to secure the transmission. A cracker can still view the message, intercept it, copy
it, and otherwise trash your communications. S/MIME is a solution to this problem.
S/MIME depends on the use of digital certificates and encryption to provide security. The
digital certificate technology relies on the use of a private key for encryption and a public
key for decryption. The sender keeps a copy of the private key, but sends a copy of public
key to anyone who needs it to read messages. Because only the sender has the private key,
only the sender can generate the encrypted message in question. The encryption process
ensures that a client can instantly detect any tinkering by a third party because a third party
won’t have the private key required to encrypt the message again. The public key makes it
easy to decrypt the message and to identify the sender during the decryption process.
Using S/MIME represents the easiest way to ensure that a server and client authenticate
each other and that the data they exchange remains safe. The public key is the weak point in
the equation. A cracker could potentially gain access to the sender’s public key (unless you
send the key by some secure method on a floppy). This means the cracker could still read
the sender’s messages and use the content, but still couldn’t generate messages on the
sender’s behalf.
Some developers consider S/MIME less secure than alternatives such as Pretty Good
Privacy (PGP) because of the way the protocol handles encryption and decryption. The fact
that the keys remain static means that the potential for a cracker to break into your system
74 Chapter 3 An Overview of Security Issues for Soap

is higher than if you used dynamically generated keys. No solution is perfect, but S/MIME
represents a relatively safe method of data transfer.

You may wonder why companies don’t rely on secure data exchange technologies like
PGP. The reason is simple—cost. Most browsers and e-mail programs understand
S/MIME without any special additions, so it’s free. You have to pay for higher security
alternatives like PGP and ensure your customers use this solution as well. Availability
and cost are why S/MIME is so popular.

Secure Socket Layer (SSL)

SSL is a protocol that’s easy to understand. As with S/MIME, SSL relies on the use of digi-
tal certificates exchanged between a client and server. A client and server obtain these digital
certificates from a third-party vendor, such as VeriSign, which can vouch for the identity of
both parties. However, unlike S/MIME, the certificate exchange procedure is a little more
complicated than simply sending a digital certificate from one machine to another. Here is
the six-step process for SSL authentication:
1. The client sends the server an unencrypted random message along with its VeriSign
issued certificate (which contains the client’s public key). VeriSign encrypted the certifi-
cate using its private key. Because everyone has VeriSign’s public key, the server can
decrypt the certificate and check it for accuracy. Also, because the certificate was
encrypted using VeriSign’s private key, no one can forge a public key of their own—
they have to get it from VeriSign.
2. Once the server confirms it’s received a valid certificate and public key from the client,
it tells the client to send an encrypted version of its original message.
3. The client computes a digest of its original random message and then encrypts it using
its private key.
4. The server uses the client’s public key to decrypt the digest.
5. The server compares the decrypted digest to a digest it generates from the random
message originally sent in unencrypted form by the client.
6. If the two digests match, the server authenticates the client.

The use of additional encryption and verification steps makes SSL slightly more secure than
S/MIME. SSL doesn’t rely on a private and public key pair, it uses a private key only. That’s
the reason this protocol requires additional handshaking steps. It’s also the reason that
crackers would have a harder time discovering a means to decrypt the message—there isn’t
any public key to steal.
SSL offers other advantages to developers. It’s transport protocol independent, which means
you can use it in many different environments. The client and server can negotiate a crypto-
graphic and encryption algorithm that both support, which means SSL is platform neutral
 User Identification Issues 75

for the most part. Even though client authentication occurs using asymmetric, public key,
cryptography, data exchange occurs using symmetric, private key, cryptography that changes
on a per session basis. Finally, the protocol is reliable—data undergoes an integrity check
before the protocol accepts it as valid.

User Identification Issues

A user name and password is normally the only form of access identification that a user
needs to gain access to your server. In most cases, that’s all you really need to ensure that the
data on your system remains secure. However, even if you make the software that grants
access to the system totally bulletproof (an ever more difficult task as computers gain
processing power), there are still problems with this approach.
Obviously, the most common problem is lost passwords. Asking the user to provide a
password that’s difficult to break often results in a password that’s also difficult to remem-
ber. Consequently, the network administrator ends up spending more time fixing lost
Ch
password problems.
Crackers and employees alike can also compromise passwords in a number of ways. The
3
least subtle scenario is the user who writes their password down on a notepad, then places
the note near the keyboard or monitor. Believe it or not, this happens on a regular basis. At
one company, I found a single notepad that contained not only the administrative assistant’s
password, but the manager’s password as well.
Even if everyone at a company does their best to protect security, and you have the best
security available for your application, crackers often find ways around the security measures
you have in place. Just the fact that your security relies on a password means that someone
can guess the password given enough time and opportunity. In other words, using the name
and password method of security is problematic at best. You can’t rely on a password to
provide total security of some types of very confidential data.
Fortunately, you have some alternatives when it comes to the password dilemma. The follow-
ing sections talk about the two most popular alternatives: smart cards and biometrics. Smart
cards rely on a something that looks like a credit card with internal memory. The network
administrator encodes the user’s password on the smart card, which the user swipes in a
special reader. Biometrics rely on unique body parts such as the iris or fingerprints for identifi-
cation. Both methods incur extra expense, but represent the best available in security today.

Using Smart Cards

One of the more common security methods today is the use of smart cards. A smart card is a
credit card-sized device with some processing power built in. The user swipes the card in a
special reader to gain access to the network. What happens is that the card provides a digital
certificate that identifies the user in lieu of a password. One of the nice things about using a
smart card is that you don’t have to provide any identification like the numbers used by
credit cards on the outside.
76 Chapter 3 An Overview of Security Issues for Soap

One company that’s starting to make smart-card readers a standard option is Hewlett-
Packard (HP). They have optimized their Vectra Desktop PCs and Kayak PC workstations
to use either the Gemplus (a Veridicom product) or Schlumberger smart-card readers. In
addition, the ProtectTools corporate security strategy offered by HP extends from the client
to the firewall, which enables a company to give outside partners secure access to company
data. The HP TopTools Management software allows managers to view clients that have
smart-card readers. In essence, there’s already an infrastructure and products in place to
allow a company to implement full security using smart cards.
A smart card does eliminate the problem of compromised passwords—at least without assis-
tance. Without a password lying around, it would take a concerted effort by an inside party
to break network security. Such an effort would be easier to trace than a break-in by an out-
side source. However, a user can still lose a smart card. Anyone who knows what the card is
used for and who it belongs to can gain access to that person’s secure resources. Unfortunately,
you still can’t stop the user from placing some form of identification on the card, even if the
issuing company doesn’t. In addition, the issuing company has to bear the cost of issuing
and maintaining the cards, which could be an expensive proposition.

Using Biometrics
There’s a relatively new alternative to both passwords and smart cards. Biometrics is a statisti-
cal method of scanning an individual’s unique characteristics to ensure that they are who they
say they are. Some of the scanned elements include voiceprints, irises, fingerprints, hands,
and facial features. The two most popular elements are irises and fingerprints because they’re
the two with which most people are familiar. The advantages of using biometrics are obvious.
Not only can’t the user lose their identifying information (at least not very easily), but also
with proper scanning techniques the identifying information can’t be compromised either.
There are two main problems with using biometrics: quality and prices. Some managers
don’t want to use biometrics because of the time it takes to create quality scans that ensure
absolute accuracy. The quality issue is being resolved as computer hardware gets faster. In
addition, the price for a single scanner can range from $100.00 for something simple to
thousands of dollars for complex solutions. Considering the number of access points a
typical company has, the price of using biometrics for all but the most stringent security
requirements could be prohibitive. Obviously, you have to weigh the cost of the data you’re
trying to protect against the cost of protecting it.
Fortunately, the tide is turning for biometrics as more companies start to pursue this solution.
Today you’ll find biometrics in use in a wide range of government and high-security institu-
tions. Many industry experts predict that forward-thinking financial organizations will begin
using biometrics identification by the end of 2002. By 2003, many corporations will view
fingerprint identification as the access method of choice for remote communication. Iris
identification may begin to overtake fingerprint identification as early as 2004 as the primary
means of identification used by some corporations.
 Where Do You Go From Here? 77

A few notable companies create biometrics solutions today. One of the most popular sys-
tems, the IriScan System 2100, can scan the iris of an individual using a camera instead of
direct contact. Find out more about IriScan, Inc. at http://www.iriscan.com/.
If your company prefers to use fingerprints instead of other biometrics techniques to
enforce security, you might want to look at the fingerprint reader chip solutions provided by
Veridicom, Inc. (http://www.veridicom.com/). They’re working with Intel to incorporate
interfaces for biometrics into the Common Data Security Architecture (CDSA). CDSA is a
comprehensive set of security services that will make secure transactions easier. It has a four
layer architecture: application, layered services and middleware, Common Security Services
Manager (CSSM) infrastructure, and security service provider modules. Find out more
about CDSA at http://www.opengroup.org/security/cdsa/.
IriScan, Inc. and Veridicom, Inc. aren’t the only companies getting into biometrics. Compaq
computer has introduced a sub $100 fingerprint reader that it developed with Identicator
Technology. You can find out about this and other Compaq biometric technologies at http://
www.compaq.com/newsroom/pr/1998/pr070798b.html. A few other biometric product dealers Ch
include Saflink (http://www.saflink.com/), Keyware Technologies (http://www.keyware.com/),
Visionics (http://www.faceit.com/), and Ethentica (http://www.ethentica.com/). All of these 3
companies provide various biometric products that can make your network just a bit more
secure from prying eyes.

Where Do You Go from Here?

Most of the security software that you can use today is vendor-specific and proprietary,
which means that you’re stuck looking at security through one vendor’s eyes. The problem
with this approach is that you may not find a single vendor who can address all of your secu-
rity concerns. Sure, most server operating systems today include a wealth of security
features, but today’s computing environment doesn’t limit itself to just one isolated company.
In a world-based Internet economy, you need the ability to secure a wide range of computer
equipment from a large number of security threats. The ability to make security solutions
from more than one vendor work together is no longer a luxury—it’s a necessity.
Security vendors see this problem as well and are working to make their software interopera-
ble, using a number of initiatives and APIs. The latest initiative to hit the market is Common
Content Inspection (CCI) API, a project started by Aventail, Corp, Finjan Software, Ltd.,
and Check Point Software Technologies. This group joins others like Adaptive Network
Security Alliance (ANSA) and Open Platform for Secure Enterprise Connectivity (OPSEC)
Alliance. The goals of these three groups vary slightly, but the main goal is to make it possi-
ble for more than one vendor’s products to work together. The largest of these groups is
OPSEC Alliance with 260 members (as of this writing).
The CCI API seeks to promote interoperability using common interfaces for inspecting con-
tent. In other words, each vendor’s software might provide completely different features and
78 Chapter 3 An Overview of Security Issues for Soap

capabilities, but they would be able to work with the same data because the means for inspect-
ing it are the same. The resulting specification is supposed to work with data at whatever level
it appears, including firewalls, antivirus checks, and mobile code. You can find out more about
CCI API at: http://www.stardust.com/cciapi/ and http://www.finjan.com/ cciapi.cfm.
OPSEC Alliance is providing a similar set of APIs through Convent Vectoring Protocol
(CVP). The difference, in this case, is that they’re concentrating on a single firewall prod-
uct, Check Point’s Firewall-1. In addition to content, OPSEC is working on APIs that allow
various vendor products to work together at the application level. At the time of this writing,
the OPSEC Alliance offering is in place and even includes a certification program. You can
find out more about OPSEC at: http://www.opsec.com/.
Like the OPSEC Alliance, ANSA caters to a specific vendor product. In this case, the target
software is Internet Security System (ISS) Group’s intrusion detection software. ANSA cur-
rently has 40 companies participating in the standardization effort. Unlike the other two
security interoperability efforts discussed in this section, ANSA offers a software development
kit (SDK) that contains Adaptive Network Security (ANS) modules designed to make inter-
operability easier. The ANS modules will allow ISS Group intrusion detection software to
issue alerts to firewalls, extranet management applications, and other software as soon as it
detects an intrusion. These other applications would then deny system access to the
intruder. You can find out more about ANSA at http://www.iss.net/.

Case Study
Sometimes security is more a matter of determined cantankerous behavior than spiffy new
technologies. Consider the case of a company that we’ll call NoGo Corporation (to protect
the guilty). The company did all of the right things. They added a firewall to their system.
The network administrator studiously applied every operating patch as soon as the vendor
found problems and issues fixes. The company had a security policy in place and held regu-
lar training sessions for users. Finally, the network administrator consistently audited the
system for unneeded user accounts and squelched them. It sounds like a formula for security
success on any network.
It shouldn’t surprise you that the company crackers invaded not once, but several times in a
two-month period. The crackers accessed the company’s database and made several thou-
sand customer accounts accessible to the public. This included the customer’s credit card
numbers and personal information such as address and telephone numbers. In short, this
fiasco cost the company a lot, even though it did everything by the rulebook.
The problem is that crackers don’t use a rulebook. They use determination and time-tested
procedures to break into computer systems. Crackers don’t care about your rulebook and
don’t intend to follow it.
The administrator for this company made the mistake of saying his system was secure during
an interview by a magazine. The magazine made a mistake by publishing the network adminis-
trator’s comment verbatim. Once the word was out on the street, crackers felt challenged and
 Case Study 79

wanted to show the world that they could break into any system. The truth is that computers
are so complex today that crackers can break into any system. A smart developer doesn’t rely
on technology alone to keep a system secure because technology is relatively easy to overcome.
In addition to saying too much, the administrator further complicated things by providing
less than stellar network monitoring. A good network administrator assumes that crackers
will get into his or her system and looks for signs of entry. A good developer provides the
administrator with monitoring tools that provide a full view of the network without imping-
ing on customer privacy. In addition, developers today need to use standardized protocols
that an administrator can update outside the application. This enables the administrator to
fix problems without getting the developer involved every time.
NoGo Corporation eventually overcame the embarrassment and loss of customers that the
security breach caused. I assume the network administrator found work elsewhere. The
company now incorporates monitoring as part of their security strategy. A few workers may
feel big brother is watching them, but customers feel more secure and that’s the bottom line.
Ch

3
 CHAPTER
4
Using SOAP to Create a Simple
Application
In this chapter
Introduction 82
An Overview of Microsoft’s SOAP Toolkit 83
An Overview of the Application 98
Shortcuts for Creating SOAP Applications Quickly 100
Understanding Namespaces, the Short Version 102
Creating the Server Side Code 104
Creating the Client Code 106
Testing the SOAP Application 109
Handling SOAP Errors 116
Performance Concerns for all Applications 117
Project 120
82 Chapter 4 Using SOAP to Create a Simple Application

Introduction
SOAP, as mentioned in previous chapters, is a specification for a wire protocol. It enables
you to transfer data between machines in a distributed application. In addition, the designers
of this specification intend to make SOAP easier to use than other wire protocols, while
overcoming some of the difficulties in using these older protocols over the Internet. We’ve
also looked at the format of a SOAP message and some of the special features you need to
consider while designing your application.
This chapter completes the theoretical part of the book and gets into some hands on exam-
ples. We’ll look at the last piece of the puzzle, the application that use SOAP to transfer data
from one point to another. I’ll present a simple example using the Microsoft SOAP Toolkit,
one of many available on the market right now.

This chapter concentrates on the Microsoft SOAP Toolkit because I feel it will be one of the
more popular toolkits on the market. However, you should be aware that there are many
SOAP toolkits available and we’ll look at some other selections as the book progresses.
You should select the SOAP toolkit that best meets your company’s needs. In some cases,
the Microsoft SOAP Toolkit is the worst possible choice (at least as of this writing). For
example, you can’t use it for PDA development. In short, there isn’t a “one size fits all”
choice in the world of SOAP today, so it pays to know which toolkits offer specific features.
I also designed the example in this chapter to fill in the gaps in your theoretical knowl-
edge of SOAP. Most SOAP applications are more complicated than the one here.

The first section of this chapter, “An Overview of Microsoft’s SOAP Toolkit,” acquaints you
with the Microsoft SOAP Toolkit. We’ll discuss toolkit features and some of the problems
that you’ll experience when using them. This toolkit is central to many of the examples in
the book, so you need to know as much as possible about it at the outset.
The second (An Overview of the Application), fifth (Creating the Server Side Code), and sev-
enth (Testing the SOAP Application) sections of the chapter contain the example application.
We’ll talk about how to put a basic SOAP application together, how the server side of the appli-
cation works, how a client requests services, and finally, how to begin testing most applications.
The third section of the chapter, “Shortcuts for Creating SOAP Applications Quickly,” pro-
vides tips you can use to get your application out the door faster. Each tip helps you get a
little more out of your development environment and produce applications with fewer bugs.
In addition, we’ll talk about ways that you can improve overall efficiency of a team develop-
ment effort.
SOAP applications rely on namespaces to find resources and external code. The fourth sec-
tion of the chapter, “Understanding Namespaces, the Short Version,” tells you how name-
spaces work within a SOAP application. This section isn’t a complete tutorial—it provides
background information we’ll build upon later. You’ll see namespaces used throughout the
rest of the book in various situations and will gain a better appreciation for them as you
work with the examples.
 An Overview of Microsoft’s SOAP Toolkit 83

No matter how well you design an application, it’s going to have bugs. Testing complexity is
a problem when working in a distributed application environment. The “Handling SOAP
Errors” section shows you how to trap errors and handle them. More importantly, it tells
you how to diagnose and squash bugs that error-handling routines report as part of their
normal operation.
Writing an application is one thing; making it run fast is another. The “Performance
Concerns for all Applications” section of the chapter provides tips for making SOAP appli-
cations run faster. This section contains general tips that you can use for any application.
Finally, the last section of the chapter will take you through the steps for a simple project of
your own. It’s important to start simply with SOAP. You’ll find that SOAP is an easy proto-
col to understand, but you need to learn tricks of the trade as part of the learning process.
This project is your first step in that learning process. It builds upon all of the other infor-
mation presented in this chapter.

An Overview of Microsoft’s SOAP Toolkit

The Microsoft SOAP Toolkit is one of many offerings on the Internet right now. My feeling is
that many developers will choose either the Microsoft SOAP Toolkit or the Apache Toolkit
(http://xml.apache.org/soap/) as their first development option because both of these com-
panies have a large following and provide a reasonable level of support. Of course, many
developers might have more than one toolkit in their arsenal and for good reason—I don’t Ch
believe any one toolkit will do everything you need for some time to come. I plan to cover
other toolkits as the book progresses, but let’s begin with the Microsoft SOAP Toolkit for now.
4
This section of the chapter discusses several important topics. We’ll begin by looking at the
features of the Microsoft SOAP Toolkit. It’s important to know what this toolkit can do for
you. A second section will tell you which types of applications you can develop using this
toolkit. The next section will look at the other side of the coin—where does the Microsoft
SOAP Toolkit fall down on the job? Finally, we’ll discuss how to install the Microsoft SOAP
Toolkit and how to use it to create an application. We won’t discuss usage in detail—detailed
information appears along with the example we create in this chapter.

This chapter relies on the Microsoft SOAP Toolkit to create all of the example code we’ll
visit. If you want to participate in the examples, you’ll need a copy of the Microsoft
SOAP Toolkit from http://msdn.microsoft.com/soap/. In some cases, you may
need more than just the Microsoft SOAP Toolkit to get the job done. You can find a
general list of Web Services resources at http://msdn.microsoft.com/
webservices/. You might want to spend some time reading about the Web Service
Description Language (WSDL) because the Microsoft SOAP Toolkit relies on WSDL at
both the client and server end of the application. The WSDL file documents the func-
tionality provided by a component, such as method calls. You can find out more about
WSDL at http://msdn.microsoft.com/xml/general/wsdl.asp.
84 Chapter 4 Using SOAP to Create a Simple Application

Toolkit Features
This section acquaints you with the Microsoft SOAP Toolkit features. Currently, the
Microsoft SOAP Toolkit doesn’t have much to offer in the way of support. You’ll find that it
includes a help file, some Visual Basic samples, WSDL generation tools, Microsoft SOAP
Messaging Object Generator Visual Basic add-in, and the DLLs you’ll need for client and
server support. It also includes some example ASP files you’ll need to provide a “listener” on
the server. (A listener is an application that runs continuously and looks for client requests.)
Future versions of the Microsoft SOAP Toolkit are supposed to contain full support for
Visual C++, including header files and example programs. Hopefully, the next edition will
also include a more complete help file as well—the current offering is sparse to say the least.
The “Problems You’ll Experience” section that follows discusses these issues in more detail.

Some of you might still have the Microsoft SOAP Toolkit version 1.0. The list of features
for version 2.0 differs significantly from Microsoft SOAP Toolkit 1.0, which relied on
ROAP.DLL and Service Description Language (SDL) instead of the current setup, which
relies on WSDL or various types of low-level access. Because of the massive changes
from version 1.0 to 2.0, the examples in this chapter won’t work with the older version
of the Microsoft SOAP Toolkit.

Let’s talk about the important pieces of the Microsoft SOAP Toolkit from an application
perspective. Table 4.1 provides a list of the DLLs provided in the Microsoft SOAP Toolkit.
This table tells how to use these DLLs within an application.

Table 4.1 DLLs provided with the Microsoft SOAP Toolkit

DLL Name Description
HLSC10.DLL Contains the HTTP Library Connector. It allows you to verify information
in the HTTP, XML, and SOAP portions of a message. For example, you
can read or write SOAP messages, or check for SOAP errors. You won’t
access this DLL directly often because most applications you create will use
the high-level application programming interface (API) access method. The
only time you need this DLL is if you want to perform a low-level SOAP
message receive and send in situations where you require absolute message
transmission control. While the low-level API access method does provide
superior flexibility, it also incurs far greater development time and more
room for error in formatting the SOAP message. This is a high-performance
connector library for use with Windows NT and Windows 2000 only.
MSSMO.DLL The Microsoft SOAP Messaging Object (SMO) Framework allows you to
work with SOAP messages as you would any other object within Visual
Basic. This DLL contains all of the classes required to work with SOAP
message parts. Using this feature helps you to create SOAP messages without
spending a lot of time learning XML. In short, the SMO Framework repre
sents the least time-intensive method for creating applications.
 An Overview of Microsoft’s SOAP Toolkit 85

Table 4.1 Continued

DLL Name Description
MSSMOGen.DLL You’ll normally use this DLL with the MSSMO.DLL. It contains the wizard
used to generate SMOs. We’ll talk more about the Microsoft SOAP
Messaging Object Generator add-in later in the chapter.
MSSOAP1.DLL This is the first SOAP DLL that you’ll use on a high-level API access client
or server. The Microsoft SOAP Library contains a number of classes, but
you’ll always begin by creating either a SoapClient or SoapServer object.
These objects support a single method call, Init(), that allows you to tell
SOAP which WSDL file to use. Other classes allow you to perform tasks
such as serializing data for output and reading input. You’ll see better how
this works later in the chapter.
WISC10.DLL This DLL contains the Windows Internet Connector Library. It provides
essentially the same services as the HTTP Library Connector described
earlier. You’ll also use this DLL for low-level SOAP message receive and
send. This lower-performance library works with Windows 98, Windows Me,
Windows NT, and Windows 2000.
WSDLGen1.DLL The WSDL generation tools use this DLL to create a WSDL based on a
server-side component you create. Normally you won’t need to access this
DLL if you use the wizard to create your WSDL file. However, it’s nice to
know that you can also create WSDL files on-the-fly if necessary by using
this DLL. It contains classes that enumerate the interfaces within a component
and write the associated information to a file. You can also examine interface
information as part of using this DLL. Ch

XHSC10.DLL This DLL contains the XML HTTP Connector Library. It provides essen- 4
tially the same services as the HTTP Library Connector described earlier.
You’ll also use this DLL for low-level SOAP message receive and send.

You’ll also find an MSSOAPR.DLL file in the resources folder for the Microsoft SOAP
Toolkit. This file contains resources for SOAP use in general and normally you won’t
need to reference it directly.

One of the first places you should look when experiencing client-side errors in your
SOAP application is the Microsoft XML Library. My workstation has three versions of
this library installed. Although the library versions are clearly marked, you can select
the wrong one by mistake. The current version of the Microsoft SOAP Toolkit relies on
version 3 of the Microsoft XML Library found in MSXML3.DLL. Microsoft is already
preparing version 4 of the same library that will appear as MSXML4.DLL.

As you can see from Table 4.1, the library that you’re going to use most often is the
Microsoft SOAP Library. It contains everything you need to begin a conversation between
client and server using the high-level API. What the table doesn’t show is that you’ll also
need the latest version of the Microsoft XML Library. This library doesn’t appear in the
86 Chapter 4 Using SOAP to Create a Simple Application

same directory as the rest of the SOAP files. The Microsoft SOAP Toolkit installation pro-
gram automatically adds the latest version of the Microsoft XML Library to your System or
System32 directory.
On those few occasions when you do need low-level API access, make sure you include one
of the three low-level API access libraries in Table 4.1. All the sample programs provided
with the Microsoft SOAP Toolkit rely on the Windows Internet Connector Library. This
doesn’t make the other selections any better or worse, it simply means you’ll spend a little
additional time figuring out how to use them. It helps that all three libraries work about the
same way, contain about the same classes with the same methods, and that you could theo-
retically use any of the three with the same boiler plate code.

Three Types of Microsoft SOAP Toolkit Application

SOAP isn’t a single application protocol. You can use it in a variety of ways and the
Microsoft SOAP Toolkit doesn’t support all of them. The toolkit does provide support for
the following three application types. (I’ve placed them in order of complexity from least to
most complex.)
■ SOAP Message Object (SMO)
■ Remote Procedure Call (RPC) (also known as High-Level API)
■ Low-Level API

All of these techniques require that you create a listener. The complexity of the listener
reflects the method you want to use for access and the complexity of the application itself.
You’ll see how listener complexity varies as the book progresses.
Two of the methods—SMO and high-level API—require the creation of WSDL files. Only
the low-level API method avoids this requirement. Fortunately, you can generate the
WSDL files with relative ease using one of two WSDL generation utilities (see the “WSDL
Generator” section for details).
The SMO method is the easiest and least flexible method of creating a SOAP application.
The client side of the picture is almost too easy. All you need to do is create an ActiveX
DLL project and use the Microsoft SOAP Message Object Generator add-in to create the
required component entries. Include this component within your client-side application
code to make accessing the remote object easy. In fact, with careful planning, neither the
server-side or the client-side code will require much change.
The high-level API technique is just a little harder than the SMO method. The Microsoft
SOAP Toolkit actually supports two forms of this project. You can use simple parameter
calls or rely on an XML document to define the interface for your application. In both
cases, the server-side component is unchanged from anything you created in the past. (You’ll
have to create a server to handle the XML document details, should you decide to use an
XML document in place of simple parameter passing.) The listener works with a WSDL
 An Overview of Microsoft’s SOAP Toolkit 87

and WSML file to create the required interface and perform the required data translations.
The client-side application does require modification to use this technique—you must make
it SOAP aware.
The low-level API technique is the most difficult to implement, yet provides the cleanest
SOAP implementation. This method also provides the greatest flexibility because you have
full control over every aspect of the communication. Both the client and server must provide
SOAP instructions. The application modules not only initiate the SOAP environment, but
create the individual lines of code passed on the wire using a serializer as well. You’ll have to
write this kind of application from scratch and you won’t be able to rely on existing code for
very much. In addition, modifications can become quite difficult because you need to
recompile the application after every change.

Problems You’ll Experience

The Microsoft SOAP Toolkit is a work in progress. For that matter, all of the toolkits that
we’ll work with in this book are works in progress, so you need to take proper precautions
when working with them. The most important issue, of course, is to ensure you keep vendor
specified limitations in mind as you work with the product. Some people are already using
these products to create production applications. Obviously, you’ll need to weigh the bene-
fits of using SOAP against the problems of using a less than stable development environ-
ment and should prepare to make changes to your code later.
Eventually, if everything works right, the output of any toolkit that you choose will be com- Ch
patible with any other toolkit. Most of these toolkits will also provide a wealth of application 4
development features—many are feature rich already. You may even find that you can even-
tually choose a single toolkit to meet all of your needs. However, there aren’t any perfect
toolkits on the market at the moment and most developers will find they need more than
one to get the job done.
One of the more interesting problems with the current implementation of the Microsoft
SOAP toolkit is that it doesn’t work particularly well with Visual C++. This is an interesting
problem because Microsoft normally emphasizes Visual C++ in its toolkits. Version 2 of the
toolkit offers little in the way of Visual C++ example code, libraries, or header files. Creating
a Visual C++ application using the Microsoft SOAP toolkit is substantially harder than other
options you may have. You may want to spend some time looking at the toolkit documenta-
tion and working with the example code before committing to using this toolkit in your next
Visual C++ development project. In fact, we’ll use the 4S4C Toolkit in Appendix C to create
a Visual C++ SOAP application.
Another problem that Microsoft plans to fix in the next version is the lack of support for
complex data types and structures. For example, you can’t pass a user defined type (UDT)
using the current version of the toolkit. (I show how to get around this problem in the
“Using Complex Data Types” section of Chapter 8.) This means that you will find yourself
looking at other toolkits when Microsoft doesn’t provide enough complex data type support
for your needs.
88 Chapter 4 Using SOAP to Create a Simple Application

You’ll want to work with Windows 2000 with Service Pack 1 installed for your early devel-
opment work. The packaging says that the Microsoft SOAP Toolkit works fine with
Windows 98, Windows ME, and Windows NT with Service Pack 6 installed, but many
people experience problems when working with these other platforms.
None of the example programs included with the Microsoft SOAP Toolkit work on a two-
machine setup. Microsoft assumes that every developer using this toolkit works on only one
machine and independently as well. Unfortunately, this is an extremely poor development
environment. While you can learn SOAP on a single machine, make sure you develop using
two or more machines to avoid many “single machine thinking” errors. I eventually modi-
fied the sample code in the Microsoft SOAP Toolkit to run on a multiple machine setup in
order to see how it would run in the real world.
The Microsoft SOAP Toolkit currently offers nothing in the way of cross-platform support
and I wouldn’t expect this to change. We have already discussed the issue of compatibility
problems with other toolkits in the “SOAP and the Web Server” section of Chapter 1, but
Microsoft intends to address this issue. You should be able to team the Microsoft SOAP
Toolkit with another toolkit such as the Apache SOAP Toolkit to provide some level of
cross-platform support. The problem, of course, is that now you’re working with two differ-
ent toolkits and will likely find implementation differences. These differences can only serve
to slow your development efforts.

Interoperability is a major concern for all SOAP toolkit vendors. However, the reality of
interoperability may take a long time to realize. You can track current interoperability
problems on the Yahoo SOAP list server at http://groups.yahoo.com/group/
soapbuilders. While this group may not solve all your interoperability problems,
they’re a step in the right direction.

PDAs are becoming more important as remote users begin using them for their daily needs.
Currently, the Microsoft SOAP Toolkit offers nothing in the way of support for PDAs. Microsoft
has committed to offering some level of support for PDAs in the future, but not with the current
toolkit release. Expect to see some form of PDA support in Microsoft SOAP Toolkit Version 3.
In the meantime, you’ll want to use another product such as pocketSOAP or IdooXoap for your
PDA needs. You’ll read about some of these toolkit alternatives in Chapter 10.
Earlier, I mentioned the requirement to use WSDL with the Microsoft SOAP Toolkit version
2.0. It turns out that the toolkit doesn’t provide full WSDL support yet, but it does provide
enough for most uses. For example, you’ll find that you can’t use the WSDL <import> tag
yet. While a future version of the Microsoft SOAP Toolkit provides full WSDL support; you
can’t use it fully now.

You can get around some of the file support problems in the Microsoft SOAP Toolkit by
using third party products such as XML Spy. You’ll find a complete discussion of XML
Spy in Appendix C, “Third Party Tool Reference.” You still need to generate the XDR
 An Overview of Microsoft’s SOAP Toolkit 89

files for an SMO application by hand or use a product such as BizTalk Server (see
Appendix B) for Version 2. Hopefully, Microsoft will add an XDR generation tool for
Version 3 of the Microsoft SOAP Toolkit.

The SOAP specification is still liquid—developers continue to suggest changes that affect
how SOAP works. Expect to see new problems pop up in Microsoft’s toolkit as the specifica-
tion changes. In addition, Microsoft is almost certain to add features to their SOAP toolkit
that will extend its functionality. In short, you’re going to run into additional compatibility
problems. How extensive the problems become is up to you. If you choose to use only the
features contained in the SOAP specification, your applications should maintain a high level
of compatibility with applications created by other developers.
A potentially crippling problem with the Microsoft SOAP Toolkit is that it doesn’t send type
information with the variables that it outputs. The reason that I say this problem is only
potentially crippling is that the SOAP specification doesn’t require vendors to provide type
information. We discussed one result of the lack of type information in an earlier chapter—
Microsoft’s solution doesn’t work well with Apache servers. However, more devastating is
the lack of support for variants. Without type information, the Microsoft SOAP Toolkit
can’t support variant types of any kind.

Ch
Many distributed applications cross human language boundaries today, which means
that you need to consider language compatibility along with other issues when devel- 4
oping an application. Consider, for example, a developer who created a SOAP applica-
tion that works fine in English only or Dutch only. However, when the English version
tries to contact the Dutch version (or vice versa), the application generates an error.
The English version transfers Boolean values as “True” and “False” while the Dutch ver-
sion uses “Waar” and “Onwaar.” The SOAP toolkit treats the two sets of values differ-
ently, making them different values as far as the SOAP application is concerned.
Another common language-related problem is the representation of numbers. Some
languages use the comma for the decimal portion of a number, while others use the
period. SOAP toolkits often experience problems translating between the two numeric
representations. A value of 99,9 in German may become 999 in English (instead of
99.9). In many cases, SOAP won’t generate an error when making an incorrect numeric
translation, but the output of the application is incorrect (which means you have to
check all cross language calculations carefully). SOAP toolkit vendors will obviously
work on these problems as they become apparent, but you need to be aware of lan-
guage issues as you develop applications.

The point of this section is that Microsoft’s SOAP Toolkit has problems. Many of these
problems will disappear with time and others will take their place. SOAP does offer the
ideal of making every application compatible; something that you really need when develop-
ing distributed applications. The reality of the SOAP world is that you’ll find a higher
degree of compatibility than ever before, but making your application work with someone
else’s application will still require cooperation from both parties.
90 Chapter 4 Using SOAP to Create a Simple Application

Using the Microsoft SOAP Toolkit

The Microsoft SOAP Toolkit contains two important tools that you need to know about.
The first is the WSDL Generator that you’ll use with every application. This tool creates an
XML description of your component that clients will use to access it properly. The second
tool is the Microsoft SOAP Messaging Object Generator Visual Basic add-in. You’ll use this
tool within the IDE to create client-side components that can interact with a server-side
component by generating SOAP messages. The advantage to using the add-in is that you
don’t need to know how to write XML.
The examples in this section of the chapter rely on a simple component that I created
specifically for this task. All this component does is add two numbers together and return a
result to the client. We’ll also use this component in the simple SOAP application example
later in the chapter. I won’t present the source code for the component in the chapter to
save space, but the source code does appear on the Que Web site for this book. You can find
it at www.quepublishing.com. You’ll find both the source code and the compiled component
in the Chapter 04/Sample Component/directory. This directory also contains all of the
ancillary files required for the application, such as the WSDL, WSML, and ASP files.
Make sure you copy the compiled version of the component to your server. Once you
have it placed in a directory that you want to use for testing, make sure you set security
to provide access to the component. We’ll look at what you’ll need to do to make the
component accessible to Internet Information Server (IIS) later in the chapter. For now,
all you need to do is register the component on both your local workstation and the
server so that you can access it. Type RegSrv32 AddIt.DLL at the command prompt and
press Enter. If the component registration works, you’ll see a success message like the
one shown in Figure 4.1.

Figure 4.1
RegSvr32 displays a
success message after
you register a compo-
nent.

You can unregister the AddIt component by typing RegSvr32 -U AddIt.DLL at the com-
mand prompt. Make sure you unregister the component before you remove it from
the server. Otherwise, the registry will continue to point to a nonexistent location and
you could receive unexpected error messages.

Now that you have some idea of what’s happening, let’s look at the two tools in the
Microsoft SOAP Toolkit. The following sections look at the two tools in detail. We’ll talk
about fixing the Microsoft SOAP Toolkit examples first, however, so that you can try them
out on two machines.
 An Overview of Microsoft’s SOAP Toolkit 91

Using the Microsoft SOAP Toolkit Examples on Two Machines

Before we go too far, let’s talk about one important issue that you’ll want to consider when
using the Microsoft SOAP Toolkit. Most developers will want to try the Microsoft SOAP
examples on two machines before they create their own projects. As stated earlier, a defi-
ciency of the current toolkit is that it assumes you want to use a single machine for devel-
opment purposes. The best way to set the Microsoft SOAP Toolkit up for two-machine use
is to install it onto a drive on the Web server from the local workstation. That way, the
installation program will add all the required SOAP toolkit entries to the local workstation,
but the files will reside on the server. Make sure you actually install all of the SOAP exam-
ples—this is an optional part of the installation process (as normal for all Microsoft devel-
opment tools).
Once you’ve installed the toolkit, create a virtual directory to the \Program Files\
MSSoapSDK\Web directory as specified in the Microsoft documentation for the samples.
You’ll also need to register all of the DLLs in the \Program Files\MSSoapSDK\Binaries
directory manually using RegSvr32 <DLL Name> on the server.
Remember that there is one special upgrade that the Microsoft SOAP Toolkit installer per-
forms that you wouldn’t spot right away. The inclusion of the Microsoft XML Parser ver-
sion 3.0 is important because your SOAP applications won’t run without it. (You’ll get
strange error messages; some of which will state the application couldn’t create an object,
but the error message will never state which object.) Of course, your workstation has the
XML Parser upgrade because you installed the Microsoft SOAP Toolkit on it, but the server Ch
may not have this feature. Make sure you copy the three XML Parser 3.0 files (MSXML3. 4
DLL, MSXML3A.DLL, and MSXML3R.DLL) to the System32 directory on your server.
You only need to register the MSXML3.DLL file using RegSvr32—the other two DLLs
provide support functionality.
At this point, your server is setup, but the Microsoft examples still won’t work. The sample
files assume that you’re using LocalHost, not a remote Web server. The fastest way to
update the code is to open each source file, look for instances of LocalHost, and replace
them with the name of your Web server. In addition, you’ll need to create new WSDL files
(found in the \Web directory) for each of the examples. The “WSDL Generator” section
that follows shows how to do this the easy way. You can check the results of your efforts by
opening one of the ASP files found in the \Web directory. Figure 4.2 shows the output of
the CalcVB.ASP file.
Notice that we’re still receiving an error message. However, it’s also important to note what
type of error message we’re receiving. In this case, the message indicates that the error
occurred internally because there’s a detail section. In addition, the error says that it couldn’t
process your request. Since we didn’t provide any input, there was nothing to process, so the
error message makes sense. This quick test shows that the server is ready to process mes-
sages. Of course, you could still run into any number of application errors. For example, the
WSDL files can contain incorrect namespace information or you can run into other prob-
lems in locating and using the component. All this test does is check the server connection.
92 Chapter 4 Using SOAP to Create a Simple Application

Figure 4.2
Check your server
setup by viewing the
results of an ASP file
request.

WSDL Generator
The Microsoft SOAP Toolkit provides WSDL support in the form of WSDL Generator. It
provides a graphical interface like the one shown in Figure 4.3. The WSDL Generator will
work with a COM component that includes a type library. The type library is essential since the
WSDL generator uses it to provide a description of the component interfaces and methods.

Figure 4.3
The WSDL Generator
utility allows you to
create WSDL files.

As you can see from Figure 4.3, the WSDL Generator begins by asking for two entries. The
first is the name of the service you want to create, while the second is the physical location
and name of the DLL that you want to access. I normally use the name of my class as the
 An Overview of Microsoft’s SOAP Toolkit 93

service name, but you can use any name you like. After you make the required entries, click
OK. If you’ve entered the right DLL location, you’ll see a Select the services you would like
to expose dialog box similar to the one shown in Figure 4.4.

Figure 4.4
Use the Select
Interface dialog to
choose which compo-
nent interface you
want to access from
the application.

Select an interface and individual methods within the interface, then click Next. WSDL
Generator will ask you to provide the URL of the listener and select a listener type as
shown in Figure 4.5. (You’ll see an example of a listener later in the chapter.) In most cases,
you’ll want to keep the XSD Schema Namespace set for the current year, unless you’re Ch
working with an older application that relies on an older specification.
4
Figure 4.5
Part of the WSDL gen-
eration process is to
select a listener type
and XSD Schema
Namespace year.

Make the listener entries, and then click Next. WSDL Generator will ask for the physical loca-
tion and name of the WSDL file as shown in Figure 4.6. This also determines the location and
name of the WSML file. Both of these files should appear in the same directory as the ASP lis-
tener (when implemented) that you create for ease of use. Notice that you can also choose
between a UTF-8 and UTF-16 file format for your WSDL and WSML files. The UTF-8 file
is easier to read with a text editor, but the UTF-16 file provides better language support.
94 Chapter 4 Using SOAP to Create a Simple Application

Figure 4.6
Determine the location
of the WSDL and
WSML files for your
service.

Select a location and file format, then click Next. You’ll see a success message stating that
WSDL Generator created the file. The only time you’ll see a failure message is if you don’t
have sufficient rights to create a file in the target directory, or some other external problem
occurs. Note that although the message only says that you’ve created a WSDL file, you’ve
also created a WSML file.

Microsoft SOAP Messaging Object Generator

The Microsoft SOAP Messaging Object Generator is a Visual Basic add-in that you can use
to create simple SOAP applications with little effort. This add-in should appear in the Add-
In Manager dialog box after you install the Microsoft SOAP Toolkit on your development
workstation (use the Add-Ins | Add-In Manager command to display the dialog box). Figure
4.7 shows an example of what your Add-In Manager dialog box will look like. All you need
to do is check Loaded/Unloaded and click OK to begin using this add-in.

Figure 4.7
A view of the
Microsoft SOAP
Messaging Object
Generator add-in
within the Add-In
Manager dialog box.
 An Overview of Microsoft’s SOAP Toolkit 95

Remember that the Microsoft SOAP Messaging Object Generator works with the ActiveX
DLL project. You’re creating a component that hides all of the interface details of a SOAP
application. The developer then uses the component as a resource in the client application
where you work with SOAP as you would with any other object. The two-part approach
does make development easier and the application more modular.
Let’s look at what you’ll need to do to create the component part of an SMO application.
The following steps will help you create a typical component.
1. Load the Microsoft SOAP Messaging Object Generator add-in if necessary. Start the
add-in using the Add-Ins | Microsoft SOAP Messaging Object Generator add-in com-
mand. You’ll see an SMO Generator—Introduction dialog box.
2. Click Next. You’ll see a SMO Generator—Specify Schema dialog box like the one
shown in Figure 4.8.

Figure 4.8
You’ll need to provide
schema information
that the SMO
Generator can use to
create the component
framework.

Ch

3. Click Browse. You’ll see an Open dialog box. The SMO Generator can theoretically
work with XML, XDR, XSD, BIZ, and WSDL files. We’ve already seen how to create
WSDL files using the WSDL Generator in the “WSDL Generator” section of the
chapter. You could also create one of the supported file types manually or use the tools
supplied with other Microsoft products such as Microsoft BizTalk Server (see Appendix
B for details). We’ll select an XDR file for this example.

For the purposes of this example I’ll use a custom XDR file that I created using XML
Spy. You’ll find this XDR file in the \Chapter 04\Sample Component directory on the
Que Web site for this book. You can find it at www.quepublishing.com.

4. Select a file and click Open. You’ll return to the SMO Generator—Specify Schema
dialog box.
96 Chapter 4 Using SOAP to Create a Simple Application

5. Click Next. You’ll see an SMO Generator—Specify Namespace dialog box like the one
shown in Figure 4.9. This is where you’ll type the Uniform Resource Name (URN) of
your organization. (See the “A Question of Globally Unique Identifiers” sidebar in
Chapter 5 for more details about URNs.) Notice that the URN shown in Figure 4.9
consists of three parts: the moniker urn, the URN actually issued to your company by a
third party, and the name of the component class. The only entry you normally need to
change is the company URN.

Figure 4.9
Entering the right URI
for your component is
essential—make sure
you enter the right
company URN.

6. Type a URI for your component, then click Next. You’ll see an SMO Generator—
Select Elements dialog box like the one shown in Figure 4.10. Note that the name and
number of elements you see will vary by the construction of the source data file. For
example, you may see different results if you use a WSDL file than if you use an XDR
file. In most cases, you’ll find that a WSDL file provides less ambiguous results than an
XDR file will.
7. Select one or more of the top-level elements, then click Next. You’ll see an SMO
Generator—Prepare Project dialog box. Note that the wizard will automatically check
the Delete “Class1” Class Module option because you don’t need this class in most
cases. Uncheck this option if you have already added global code to the Class1 module.
8. Click Next. You’ll see an SMO Generator—Change Class Names dialog box as shown
in Figure 4.11. Note that this dialog box allows you to change the name of the classes
that the Microsoft SOAP Messaging Object Generator creates. Depending on how you
generate the source file and the names chosen by any tools you use, you may need to
change the class name to make it more readable. In this case, I changed the class name
to AddItAccess because that name is a lot more specific than Root, the name found in
the XDR file I used.
 An Overview of Microsoft’s SOAP Toolkit 97

Figure 4.10
After you’ve entered
all of the required pre-
liminary information,
you’ll see a list of
available elements.

Figure 4.11
Use an easy to deci-
pher class name, even Ch
if this means changing
the default entry 4
shown in this dialog
box.

9. Click Next. You’ll see an SMO Generator—Finish dialog box.

10. Click Finish. The Microsoft SOAP Messaging Object Generator will automatically
create all of the required classes for you.
98 Chapter 4 Using SOAP to Create a Simple Application

At this point, the wizard has completed its work. Depending on how well the source file
(component description) defined your component, you should be able to compile this DLL
without any change and begin using it for your client. All you’ll need to do is include this
component and begin programming your SOAP application using standard dot syntax.

An Overview of the Application

The example application is going to be a simple server-side component that adds two num-
bers together. The client will provide two integers as input and expect an integer as output.
In other words, this application won’t do anything amazing. My intent in creating it is to
show you some of the basics of creating a SOAP application using the simplest means possi-
ble. Rather than wade through reams of source code, you’ll see SOAP in its simplest form.
The purpose, of course, is to make it easier to understand the harder applications that
appear later in the book.
The following sections will help you understand various SOAP application elements. The
first section shows how differences in SOAP application requirements will affect your pro-
gramming methodology. Next, I’ll show you how SOAP applications exchange data in
general. Finally, we’ll look at the data flow for this application in particular.

How SOAP Applications Differ

It’s important to use the right techniques when creating a SOAP application because using
the wrong technique can result in many wasted hours reinventing the wheel. In general,
SOAP applications differ in several ways. Here’s a list of what I consider the most impor-
tant criteria.
■ Application Complexity: As application complexity increases, so do development prob-
lems. You may need to rely on something other than simple parameter passing methods
when working with complex applications.
■ SOAP Toolkit Limitations: We talked about the kinds of applications you can create
with the Microsoft SOAP Toolkit in the “Three Types of Microsoft SOAP Toolkit
Application” section of the chapter. The Microsoft SOAP Toolkit won’t provide every
solution you need. You’ll definitely want to look at all of the available solutions when
designing a SOAP application.
■ Platform: If your SOAP application has to operate on more than one platform, you’ll
need to find one or more toolkits that work with each other as part of your develop-
ment solution. For example, the Microsoft SOAP Toolkit doesn’t work with an Apache
server, so you’ll need to find an Apache specific solution if you need to work with both
Microsoft and Apache servers. Likewise, PDAs all require individual toolkits—there
aren’t any solutions that work with all PDAs.
■ Access: The type of resources your application requires affects the way you develop it.
Some developers feel that SOAP is ill equipped to handle any form of database applica-
tion. Others feel that SOAP isn’t secure enough for sensitive data. It’s important to con-
sider the resources your application requires before you decide on a design method.
 An Overview of the Application 99

■ Client Needs: The client doesn’t affect your SOAP application as much as it affects
binary solutions, but problems remain. You need to consider how the client interacts
with the server because some clients require more input than others. For example, the
infamous data-type problem affects clients as well as servers.
■ Partnerships and Customers: One of the reasons that people are excited about SOAP is
that it allows easy access to company resources by partners and customers. Of course,
this access also presents challenges to the developer. You need to consider additional
security and other problems that will affect your application design.
■ User Needs: Applications today have to support all kinds of users. Some users require a
PDA, while others can get by with a laptop computer. You’ll also need to consider the
special needs of users with special challenges, such as the visually impaired.

This list contains the issues you’ll need to consider most often. You may also have to con-
sider special needs for your company. For example, some applications will need real-time (or
nearly real-time) access to data. This requirement affects the way you build an application
and could affect your choice of tools.
The main point of this section is that SOAP actually increases the diversity of applications
you can create. No, you won’t create a desktop application with SOAP, but SOAP applica-
tions could contain some of the same functionality that used to appear with desktop applica-
tions alone. For example, administrators used to worry only about LANs. Now they need to
consider organizations with a worldwide scope. The desktop application they used to manage Ch
the network no longer fits the bill—they require a distributed application of the same type.
4
SOAP applications vary in so many ways that careful design will now include walkthroughs
and careful research. Product decisions are more critical than ever before—a poor choice
will cost more. Security hazards are greater; user requirements are greater than at any other
point in history. SOAP is a tool with a lot of potential to solve the significant problems you
face today.

Basic Application Design and Data Flow

Some people will have the wrong view of SOAP from the outset because prior development
experiences and the features of their toolkit of choice influence them. Some developers view
SOAP today as a two-way, real-time, synchronous protocol. This is the view that many
SOAP toolkits espouse today because SOAP is still in its infancy and there are few specifica-
tions to guide developers. The SOAP specification doesn’t contain any such limitations.
A SOAP communication is a one way protocol. A client sends a message to a server. There’s
no need for the server to respond to the request, nor does the server have to handle the
message on any deadline. This simple view of SOAP leaves a lot of room for innovation and
some developers are already stepping up to the challenge.
Future SOAP toolkits will very likely allow asynchronous one way communication. These
toolkits may offer queued message transmission (Microsoft is already working on this fea-
ture) and some toolkits may offer transactional support. If you’re looking at SOAP in a
specific way today, be prepared to see things differently tomorrow.
100 Chapter 4 Using SOAP to Create a Simple Application

How does the reality of SOAP affect application design and data flow? You normally need to
design a SOAP application with the one-way transmission in mind. A request/response sce-
nario, therefore, actually requires two one-way transmissions. The first sends a request to the
server, while the second sends a response to the client. Both server and client require a serial-
izer to create the message, a listener to receive the message, and a parser to read the message.
SOAP is a stateless protocol. This means that you can’t depend on the component having
any given state when you send a call. You must create a new connection each time you want
to communicate with the server because the Web server will consider each call a new ses-
sion. As you can see, making multiple calls for one piece of data is a difficult and error prone
task you should avoid. It pays to make SOAP applications that use self-contained calls.

Understanding the Simple Application in This Chapter

The application in this chapter is a perfect example of the self-contained call. We’re perform-
ing a simple task—adding two numbers. While this isn’t representative of the real world
applications that you’ll create, it does show how to use simple parameter passing within an
application. Simple parameter passing represents the easiest SOAP application you can cre-
ate. Here is the list of steps the application follows when processing the client request.
1. The user enters application data and clicks Do It.
2. The client application creates a session with the server by creating a SOAP client and
specifying the name of a WSDL file to receive the input.
3. The XML Parser on the server opens the WSDL file and processes it. This creates a
connection between the WSDL file and the ASP used to receive input.
4. The ASP receives input from the client and creates the SOAP server required to
process it. The SOAP server uses the schema found in the WSDL and WSML files to
create the required component and make the client request.
5. The SOAP server creates a response based on the output from the component.
6. The XML Parser creates a SOAP response message, which it returns to the ASP.
7. The Web server responds to the new ASP information by sending a response back to
the client application.
8. The client application parses the information and displays the result for the user.

Shortcuts for Creating SOAP

Applications Quickly
This section contains tips that help you develop SOAP applications faster and with fewer
bugs. The bugs portion of the picture is extremely important because it doesn’t pay to
develop an application with lots of bugs. You need to work fast and develop an application
that works properly at the same time in today’s market. You might find that not all of the
tips mentioned here are applicable to you project. I’ve provided tips for many different
working environments and you may work in only a few of them.
 Shortcuts for Creating SOAP Applications Quickly 101

The most important tip that I can pass along is to work with a SOAP toolkit. Yes, you can
write an application without relying on a SOAP toolkit, but it will take a lot longer and have
more bugs in the end. A SOAP toolkit provides you with tools that you need to create appli-
cations quickly, ensures the syntax of your SOAP messages are correct (and follows the stan-
dards specifications), and provides you with some “boiler plate” code such as listeners. Every
SOAP application requires a listener on the server to respond to client requests, so this is an
important part of any application.
Use the default serializer whenever possible. SOAP relies on an XML formatted text mes-
sage to transfer data from one point to another. This means that the data must follow the
current SOAP and XML specifications to work. In most cases, the rigid structure of the
XML message leaves little room for creativity when it comes to the message format (the
content of the message is a different story). Nevertheless, some developers are already creat-
ing custom serializers that perform special tasks. Using a custom serializer opens your appli-
cation to bugs and compatibility problems. The vendor creating the default serializer for the
toolkit you use has already extensively tested it against the specification, so using it when-
ever possible eliminates one additional source of problems.
Always test your application on two machines, preferably separated by an Internet connec-
tion. Some vendors continue to suggest developing an application on a single machine to
reduce the amount of remote debugging a developer needs to perform. Using a local imple-
mentation of a distributed application almost never works. The resulting application is always
going to have problems. Developing in a distributed environment from the outset brings Ch
these problems to light early in the development cycle when it’s less expensive in time to fix
them. In addition, working with application pieces on local machines in a team environment
4
will almost certainly result in problems when you put the application together. Local applica-
tion development may be easier, but it’s also one sure way to create problems in the end.
Most developers know that they need to design the data portion of an application first,
which includes creating any required databases, indexes, and views. However, when using
SOAP, you also need to design the message format of your application as part of the data
design process. Make sure you create a design that will work with partner applications if
your distributed application will affect more than one company. When a company already
has a message format in place that doesn’t agree with your own design needs, then you’ll
have to rely on technologies such as eXtensible Stylesheet Language Transformations
(XSLT). Using XSLT allows you to transition between message formats with a minimum of
programming. Of course, you can always write an application to perform the transition as
well, but this is a labor-intensive method with dubious benefits for the developer.

Like everything else dealing with SOAP, you’ll find that XSLT is in a state of chaos as stan-
dards committees hammer out the final details of a specification. You can find out more
about the XSLT specification at http://www.w3.org/TR/xslt. Find out about the
eXtensible Stylesheet Language (XSL) at http://www.w3.org/TR/xsl/. Some inter-
esting related technologies include the XML Information Set (http://www.w3.org/
TR/xml-infoset/) and XML Path Language (http://www.w3.org/TR/xpath).
102 Chapter 4 Using SOAP to Create a Simple Application

We’ll discuss most of these technologies as the book progresses, but it pays to do some
research up front. Fortunately, you don’t have to become a guru in any of these tech-
nologies to work with SOAP and use the sample applications in the book.

Use the wizards whenever possible. The WSDL Generator makes short work of creating a
WSDL file for your component. In fact, I wouldn’t consider creating an application without
this handy utility. Using the SMO Framework and associated Microsoft SOAP Messaging
Object Generator add-in for Visual Basic makes client application easier. Of course, like any
use of a wizard, the ones you find in the Microsoft SOAP Toolkit do tend to hide imple-
mentation details and are somewhat rigid when it comes to developing an application. If you
really need to create a low-level API application to maintain full control over the environ-
ment, then be sure to allow a lot of additional development time.

Understanding Namespaces, the Short Version

Libraries, DLLs, and other forms of external code storage have always presented problems
because finding the particular method or function that you need is time consuming. The
method and function declarations appear to have no discernable order and you might find
two related functions in different DLLs. In short, using libraries and DLLs in the past
required the developer to memorize the location of certain functions. While many develop-
ers did end up creating their own “internal library” of functions and their location, the
process for creating this library is expensive in terms of time.
Namespaces provide a means of bringing order to the world of library and DLL method
storage. Yes, the functions and methods of the past are still stored in libraries and DLLs, but
now you can find them much easier. In addition, you normally don’t need to memorize their
location. You can find the methods and functions with relative ease using a browser or other
tool that relies on the namespace hierarchy to do its job.
We’ve already seen how SOAP uses namespaces in the examples in Chapters 1 and 2.
Applications such as Windows Explorer use the namespace concept to organize objects like
drives and file folders. Namespace extensions provide methods of interacting with these
objects. Future versions of Visual Studio will also rely on namespaces. The Common
Language Runtime (CLR) that Microsoft plans to introduce with Visual Studio.NET relies on
namespaces to organize methods. Figure 4.12 provides you with a glimpse of how CLR will
eventually present functions for your use. In short, namespaces are a great organizational tool.
You can view all namespaces as a hierarchical database. A top-level container holds one or
more object containers. Each object container can hold one or more methods and one or
more additional object containers. The hierarchy continues until the entire content of a
DLL appears somewhere within the hierarchy. Look again at Figure 4.12 and you’ll see the
hierarchical organization of objects and methods. The objects appear in the left pane, while
the methods associated with the selected object appear on the right. Vendors always orga-
nize namespaces in this manner, but you won’t always see them displayed as shown in Figure
4.120. The tool you use for viewing a namespace determines the method of presentation.
 Understanding Namespaces, the Short Version 103

You gain access to a particular method by presenting the path to its location in the name-
space hierarchy. Look again at Figure 4.12 and you’ll see that you can access the
CloseHandle() function by requesting it like this:

Microsoft.Win32.Win32Native.CloseHandle()

Figure 4.12
CLR will rely on name-
spaces to organize
methods for developers.

Ch

4
Some application environments use a dot syntax as shown above, or they can use other char-
acters such as the colon (:) that SOAP relies on. In all cases, you need to provide an entire
path to access the desired method. Most vendors provide some type of shortcut for accessing
the path such as the Imports statement used by Visual Basic.NET. Once you define a path to
the desired method, all you need to do is reference it by name within your code, which
greatly reduces the amount of typing required to create an application.
Let’s get back to SOAP namespaces. There are two sets of namespaces to consider. The first
set appears within the SOAP message. You’ve already seen these namespaces in action in the
examples in previous chapters. The SOAP toolkit vendor defines the second set of name-
spaces that you’ll use within your application.
As of this writing, most SOAP toolkit vendors provide two namespaces as part of the toolkit.
The first namespace is for the SOAP serializer (a component that organizes the data you
want to output into the SOAP message format). The serializer namespace contains all the
methods you’ll use to create requests and responses. For example, a startBody() method call
creates the message body, while an endBody() method call ends it. The second namespace is
for the SOAP fault mechanism. It contains methods for creating the four SOAP fault ele-
ments that we discussed in the “SOAP Fault Messages” section of Chapter 2.
You’ll rely on two sources for the SOAP message namespaces. The first source is defined by
standard organizations to provide standard SOAP message features. For example, the
104 Chapter 4 Using SOAP to Create a Simple Application

http://schemas.xmlsoap.org/soap/envelope/ Web site defines standard features for the

SOAP envelope. Figure 4.13 shows an example of the fault namespace as defined by the
standards organization at the time of writing.

Figure 4.13
The SOAP specification
defines sources for
standard SOAP name-
space definitions.

You’ll also define namespaces of your own that define access to components on the server.
The SOAP toolkit usually provides a utility to create the files required to define the name-
spaces you’ll access from within your application. We looked at the process for using a cus-
tom namespace as part of the application code within Chapters 1 and 2. Essentially, you
define a local variable that references the Web site with the namespace definition. The local
variable then serves as the path to the component methods that you want to access.
This is obviously a quick tour of namespaces for SOAP, but we’ll review this topic often as
the book progresses. You can’t create a SOAP application without knowing something about
namespaces. I feel this is a topic best learned by doing, so I’ll expand your knowledge of
namespaces in the example applications that follow.

Creating the Server Side Code

You’ll normally design and test the server side code first. The reason is simple—you need a
functional server before you can do anything with the client. There are many ways to design
the server-side code, but I normally begin with the component. That way I can verify opera-
tion locally and ensure the rest of my server code won’t suffer from a faulty component dur-
ing the debugging process.
Creating a listener comes next. We’ll use an ASP file in this case. The ASP file should con-
tain code that passes input from the client to the SOAP sever and then passes the output
from the SOAP server back to the client. Most listeners will also contain some form of error
handling code. This code creates a SOAP fault message and passes it back to the client.
The final piece is creating the required WSDL and WSML files. We’ve already performed
this step in the “WSDL Generator” section of the chapter, so I won’t repeat that step here.
You’ll want to verify the server setup, at this point, by running a small test discussed in the
“Designing a Listener” section.

Designing the Component

The server-side component for this example could be any component that you created in
the past. Nothing that you’ve done in the past will change. The server-side component is
 Creating the Server Side Code 105

totally unaware of any interaction with SOAP because of the way that SOAP calls on the
component for services. Here’s the code for our component.
Public Function DoAdd(Add1 As Integer, Add2 As Integer) As Integer
‘Calculate the result.
DoAdd = Add1 + Add2
End Function

All you need to do is create a new ActiveX DLL, add a function to it, and add this code.
You’ll need to check the Unattended Execution and Upgrade ActiveX Controls options in the
AddIt—Project Properties dialog box. Make sure you set this up as a single-threaded DLL.
Compile the code as a DLL. You can test the component locally if you want, using a small
test application. Move the component, the server, and register it using RegSvr32 when you
finish. Obviously, most components are more complex than the one shown here, but the
process you’ll follow is essentially the same in all cases.

Designing a Listener
Microsoft doesn’t provide any automated tools for creating a listener. However, you can
use the listeners provided in the Microsoft SOAP Toolkit as examples of what you need to
do. In fact, these listeners will work with simple applications with just a few modifications
on your part. A simplified form of the ASP file we’ll use follows. This version doesn’t
include error handling. (I’ll talk about error handling in the “Handling SOAP Errors”
section of the chapter.) Ch
<%
4
Dim SoapServer
Dim WSDLFilePath
Dim WSMLFilePath

Response.ContentType = “text/xml”

Set SoapServer = Application(“AddItServer”)

If SoapServer Is Nothing Then

Set SoapServer = Server.CreateObject(“MSSOAP.SoapServer”)

WSDLFilePath = Server.MapPath(“AddIt.wsdl”)
WSMLFilePath = Server.MapPath(“AddIt.wsml”)
SoapServer.Init WSDLFilePath, WSMLFilePath

Set Application(“AddItServer”) = SoapServer

End If

SoapServer.SoapInvoke Request, Response

%>

As you can see, the listener is relatively simple. It begins by creating a partial response mes-
sage that states the type of content the server will return. The ASP then creates a SOAP
application (if necessary), maps a path to the WSDL and WSML files, and finally initializes
the SOAP server. The last step is to invoke the SOAP server with the request information
and pass the response back to the client. SOAP listeners are normally more complicated
106 Chapter 4 Using SOAP to Create a Simple Application

than this, but not by much. All that the SOAP listener does is pass information to the
appropriate place.
Once you have the listener, WSDL, WSML, and component in place, you can run a quick
test to ensure that you’ve set your server up correctly. Simply log onto the Web server and
select the ASP page. You should receive an error message stating that you haven’t provided
input variables like the one shown in Figure 4.14 shows the error message you should
receive. Notice that the error reflects an inability to load the XML parser (MSXML).

Figure 4.14
Perform a quick test
on the server installa-
tion to ensure you
have the application
set up correctly.

Creating the Client Code

You’ll begin creating the client by starting a new Standard EXE project in Visual Basic.
Figure 4.15 shows the form that you’ll create. Table 4.2 provides settings for this form. As
you can see, it uses a simple form that contains just enough entries to show the input and
output of the component, and allows the user to initiate the SOAP call.

Table 4.2 AddIt Client Application Form Settings

Control Property Setting
frmMain BorderStyle Fixed Single
Caption AddIt Example
Height 2000
 Creating the Client Code 107

Table 4.2 Continued

Control Property Setting
Width 5200
txtAdd1 TabIndex 1
Text 10
txtAdd2 TabIndex 2
Text 5
txtResult TabIndex 5
cmdDoIt Caption Do It
TabIndex 3
cmdQuit Caption Quit
TabIndex 4

Figure 4.15
This is the simple form
we’ll use for the client-
side application.

Ch

The client-side code needs to know about the SOAP connection when employing the tech-
nique used in this application. However, the client-side programming requirements are
light. The simple client code used for this example follows.
Option Explicit

Private Sub cmdDoIt_Click()

‘Create the SOAP client.
Dim Client As SoapClient
108 Chapter 4 Using SOAP to Create a Simple Application

‘Set up an error handler.

On Error GoTo ErrorHandler

‘Create the connection.

Set Client = New SoapClient
Client.mssoapinit _
“http://winserver/soapexamples/AddIt.wsdl”, _
“AddIt”, _
“AddItPortType”

‘Perform the addition.

txtResult.Text = CStr(Client.DoAdd(CInt(txtAdd1.Text), _
CInt(txtAdd2.Text)))

Exit Sub

‘Display a message when an error occurs.

ErrorHandler:

MsgBox Client.faultstring, vbExclamation

End Sub

Private Sub cmdQuit_Click()

‘Exit the application
End
End Sub

The example begins by creating a SOAP client. The SOAP client creates the connection
between the client and the server. You call the mssoapinit() method and supply a minimum
of one argument, the location of the WSDL file. Our example also supplies a service name
and the name of the service port. The SOAP server uses the first service and port number if
you don’t supply these values. The mssoapinit() method also accepts a fourth argument—
the location of the WSML file. Never supply this value unless you want to provide custom
schema handling.

The arguments for the mssoapinit() method call get passed to the server precisely
the way you type them into the application. If a server provides case-sensitive handling
of data from the client, then the call will fail if you don’t use the right case. It always
pays to use the same case for these arguments as you find in the WSML file.

So, where do you get the service name and service port number? The best place to look is in
the WSML file. Here is the WSML file for this example.
<?xml version=’1.0’ encoding=’UTF-16’ ?>
<!-- Generated 03/03/01 by Microsoft SOAP SDK WSDL File Generator, Version 1.0 -->
<servicemapping name=’AddIt’>
<service name=’AddIt’>
<using PROGID=’AddIt.AddItClass’ cachable=’0’ ID=’AddItObject’ />
<port name=’AddItPortType’>
<operation name=’DoAdd’>
<execute uses=’AddItObject’ method=’DoAdd’ dispID=’1610809344’>
 Testing the SOAP Application 109

<parameter callIndex=’1’ name=’Add1’ elementName=’Add1’ />

<parameter callIndex=’2’ name=’Add2’ elementName=’Add2’ />
<parameter callIndex=’-1’ name=’retval’ elementName=’Result’ />
</execute>
</operation>
</port>
</service>
</servicemapping>

Notice the <service name> and <port name> tags on the fourth and sixth lines. These two
tags define the names that you’ll provide within the client application. This is another rea-
son to generate the WSDL and WSML files before you begin writing the client application.
The WSML file will always contain the information you need and in the proper case.
Now that we have a connection to the server, let’s get back to the remaining lines in the
client application. The next step is to call the DoAdd() method within the component. This
line may look a little confusing because of all the conversions taking place. Essentially, we’re
making a straight call to the component, just as you would for any other application. The
remaining lines of code in the listing provide simplistic error handling and a means to exit
the application.

Testing the SOAP Application

Developing SOAP applications means learning new techniques. I’m sure that most of you
Ch
didn’t have to worry about using a serializer with your LAN applications, but it’s an essential
part of a SOAP application. Fortunately, once you know what a SOAP message should look 4
like, creating one isn’t as hard as it might first appear. The programming is relatively simple,
in fact, because the vendors creating the SOAP specification went to extremes to make sim-
ple code a reality.
Testing, however, will never get easier. I remember thinking that testing applications on the
desktop was difficult. Moving to the LAN brought an unprecedented level of complexity,
but it was still manageable. Testing a SOAP application, unfortunately, means checking
things that you never thought to test before. No longer can you assume anything about the
application environment. A SOAP application that works on your local machine is no longer
enough. Even testing across the LAN in your company is no longer enough because you
never know where your application will end up in the future.
The following sections are going to try to take some of the pain out of SOAP application
testing for you. We’ll begin by looking at how you can test the simple application we’ve cre-
ated in this chapter. The hands on time you spend with simple applications at the outset will
greatly reduce the learning curve for complex application in the future. The next section
will provide you with some tips about SOAP application testing. You’ll want to read this sec-
tion carefully because distributed applications present problems that could become terrifying
in scope. Finally, we’ll look at an interesting Web site you can visit to test your server for
SOAP compatibility. Testing your server against a standard is important because it ensures
that your application will work with the broadest range of external servers possible.
110 Chapter 4 Using SOAP to Create a Simple Application

Checking the Example Application for Errors

At this point, you can run the client application. If you enter two numbers, click Do It, and
get a result back from the server, everything is working fine. This is one case when you’ll
want to test the application both inside the IDE and out. The IDE can actually interfere with
the operation of your application in some cases, especially when you’re in a debugging mode.

One of the more common problems new developers run into when trying to get an
example to run on two machines is not registering the components on the server.
Installing the Microsoft SOAP Toolkit automatically registers the required components
on the local workstation, but does nothing for the server. You must register all of the
DLLs found in the Binaries directory using RegSvr32 <DLL Name>.

If you do run into problems, the first thing you’ll want to do is check the client to ensure it’s
transferring data to the server in the correct format. I’ll show you how to do this using the
tcpTrace tool in Appendix C. However, there are other resources at your disposal, such as
network monitor.
Make sure you check your application from multiple machines and offsite if possible.
Sometimes you can’t check your system from an offsite location with ease. That’s where the
SOAP Message Builder (http://www.soapclient.com/soapmsg.html) comes into play. Figure
4.16 shows what this site looks like.

Figure 4.16
The SOAP Message
Builder allows you to
simulate outside
access of your Web
site.
 Testing the SOAP Application 111

As you can see, you enter your server address, SOAP action, and SOAP message as a mini-
mum. The SOAP message should contain the same information that you expect your client
to send. You can obtain the message from the tcpTrace diagnostic screen. It’s just a matter of
cut and paste. Just click Execute when you’re ready to test your application.
The Web site also provides a sample that you can generate by clicking Generate at the bot-
tom of the page (not shown in Figure 4.16). The sample gives you a better idea of how this
Web site works. The result is that you should get the same result from your server using an
external source as you would from an internal source, if external access is an important ele-
ment of your application. Likewise, you can also use this site to test security. Given a valid
message and other parameters, this site can answer the question of whether someone outside
your company can access your application with ease.
The Generic SOAP Client (http://www.soapclient.com/soaptest.html) is another valuable
testing tool. In this case, you make a call to your server just as you did with the client appli-
cation earlier. However, in this case you’re checking external access. Figure 4.17 shows what
this tool looks like.

Figure 4.17
The SOAP Client
Builder allows you to
simulate outside
access of your Web
site. Ch

As you can see, all you need to enter is the URL for the WSDL file on your server and
the intended response type. Click Retrieve when you’re ready to perform the test. The
testing tool takes care of the rest for you. Like the SOAP Message Builder, this tool pro-
vides a simulation of external access and can point out potential security problems with
your setup. Note that this site also provides a sample application that you can use for
learning purposes.
112 Chapter 4 Using SOAP to Create a Simple Application

SOAP Testing Tips

Distributed application development represents a new obstacle to testing. Consider for a
moment that the application isn’t on the desktop or even the LAN anymore. The applica-
tion communicates across a network of unknown quality and there’s a potential for problems
that you never considered in the past. In addition to the problems of communication, a dis-
tributed application today may execute on more than one platform. In other words, the
operating system is no longer a constant, so you can’t depend on any particular set of fea-
tures on the other end of the data connection. Finally, you can’t even depend on a given
level of user training. Users from all over the world could end up using your application. In
short, the development environment for all developers today is more complex than the com-
mercial (shrink-wrap) development environment of yesterday.
Since today’s developer can’t depend on a specific environment, testing takes on a nightmar-
ish appearance at times. Newsgroups now contain impassioned pleas from developers who
have no clue at all on how to debug a remote application. In many cases, the testing proce-
dure is arcane and relies on outdated techniques, such as sending test messages back and
forth to see if you can get anything to work. With this in mind, let’s look at some simple tips
for reducing test complexity.
■ Always test your application on a LAN using two machines at the outset. Yes, I’ve
said this before, but it’s an important issue. Creating an application on the local
machine and testing it there results in an application that runs fine locally. Microsoft
and other vendors espouse this dangerous technique because it makes it easier for devel-
opers to feel they’ve accomplished something more quickly. In the end, however, a
developer only hurts the final product by testing locally.
■ Once you get an application running on LAN, test it across an Internet connec-
tion, even if you have to set up a test site from your home to the office. Your
LAN is a safe environment where messages are unlikely to get lost and the connection
is sturdy. The Internet provides none of these things. If you deploy a distributed appli-
cation on the Internet with only LAN testing under your belt, be prepared to get lots of
support calls from irate users.
■ Test your application against standards validation sites. We’ll talk about a site to
test your server in the “Testing Your Server” section. In past chapters, I’ve mentioned
two important Microsoft SOAP test Web sites http://www.soaptoolkit.com/
soapvalidator/ and http://www.soaptoolkit.com/soapvalidator/listener.asp. These
sites are important tools in a world of SOAP development where few tools exist.
■ Check your application against more than one toolkit. The compatibility problems
between the Microsoft and the Apache toolkits demonstrate the need to perform this
simple check. These toolkits won’t talk to each other at present because of the way in
which they implement the SOAP specification. Sure, Apache intends to fix the problem,
but you’ll likely run into other problems later.
■ Run your application on more than one Web server if possible. SOAP applications
depend on the features provided by Web servers to a certain extent. Make sure you
 Testing the SOAP Application 113

check your application against more than one Web server. In addition, make sure you
let users know which Web servers you checked so they know in advance that the appli-
cation may not work with other Web servers.
■ Run your application on more than one platform if possible. Distributed applica-
tions typically need to communicate with multiple platforms. You may use all Windows
2000 Server setups on Intel machines in your company, but a partner may rely on
UNIX servers that run on some other processor. It’s important to consider every poten-
tial obstacle and this is likely to be a big problem.

You’ll also find that testing your application once is probably not enough. Validating your
application every time a new specification comes out may become a painful reality in the
world of distributed applications. Unlike the application you run on a LAN, SOAP applica-
tions won’t run in a vacuum. Business partners may decide to update their SOAP implemen-
tations outside the normal processes for your company—making it possible that an
application that worked fine yesterday will no longer work today.
By now you’re thinking that SOAP is supposed to make things simpler, but this sure seems
more complicated than anything you did in the past. The truth is that SOAP has made the
development process a lot easier, but it looks more complex because of the increase in envi-
ronment complexity. Imagine trying to make distributed applications of the kind that we’re
talking about here work with DCOM or CORBA. You’d find yourself pulling out hair by
the handfuls and never getting anything done. SOAP makes distributed applications possi- Ch
ble using a standardized specification, something that all developers need in today’s con-
nected environment. 4

You’ll discover that an important part of the troubleshooting process is to know pre-
cisely what information the client and server exchange. The only way to do this is to
read this information after it leaves all of the processing levels of both machines. Since
this is such an important requirement, you’ll want to add an easy to use tracing tool to
your arsenal. One such tool is tcpTrace. We’ll discuss tcpTrace in detail in Appendix C.

Testing Your Server

Testing your server is every bit as important as testing the rest of the application. You need
to know that the server is compliant with the current SOAP standard. The problem is that
you can’t really test your server without knowing everything there is to know about the
specification. Once you were “all knowing” about the specification, you’d have to write a
test suite, and then test your server—a lot of work for a test you’ll perform once.
Fortunately, there’s another solution. You can use the SOAP 1.1 Validator site at
http://validator.soapware.org/. Figure 4.18 shows the opening screen for this Web
site. As you can see, the site generates scripts that it uses to test your server for SOAP
compatibility.
114 Chapter 4 Using SOAP to Create a Simple Application

Figure 4.18
The SOAP 1.1 Validator
Web site can test your
server for SOAP 1.1
compatibility.

Using the SOAP 1.1 Validator is easy. As shown in Figure 4.19, enter the URL for your site,
port number for HTTP access, and path to the test area. Notice that this site includes a
default entry. I’ve used this entry to learn more about SOAP by seeing how the default site
reacts to the various scripts. You can display the various SOAP messages used in the test.
This allows you to see ho’w the request/response mechanism should work and can help you
locate problems with your own site.

Figure 4.19
Using the SOAP 1.1
Validator is easy—just
enter the required site
information.
 Testing the SOAP Application 115

Once you enter the required information, click Submit. Be patient as the site tests your Web
site. I often wondered if something had gone wrong, only to see the test results appear a few
seconds later. During my tests, validation can take up to five minutes. The test time varies as
a function of your connection speed and the load on the remote server. Figure 4.20 shows
the type of results that you should get from the test.

Figure 4.20
The SOAP 1.1 Validator
provides a simple
pass/fail report of your
server’s compatibility.

The Web site will display several columns of information. The second column contains a
description of the test. Click on the link to gain insights on test purpose and server response Ch
requirements. The third column displays OK if your server passed the test or a failure mes-
sage if it didn’t. The fourth and fifth columns are the most interesting. They help you to see 4
the test site request and your server’s response. You can use this information for both train-
ing and diagnostic purposes. It could also come in handy for testing the capabilities of vari-
ous toolkits. Figure 4.21 shows a typical response message.

Figure 4.21
A typical server
response to a SOAP 1.1
Validator request.

As you can see, this is a textbook response similar in format to those we discussed in
Chapters 1 and 2. This particular response is plain—it doesn’t include many of the attributes
we discussed earlier, but it does meet the minimum SOAP requirements. Obviously, you’ll
116 Chapter 4 Using SOAP to Create a Simple Application

want to check your server’s response against the level of SOAP message response that your
application actually requires. For example, a response should include embedded type infor-
mation if you plan to work with complex data types.

Handling SOAP Errors

SOAP errors come in a variety of shapes and sizes. You’ll find that debugging a SOAP appli-
cation is every bit as complex as any other distributed application, especially if you don’t
have the right tools. For example, finding a problem between two platforms becomes quite
complex when you need to use two different toolkits. Compatibility problems are at the
forefront of SOAP issues today.
Many of the most frustrating SOAP problems are also the easiest to detect and fix.
Developers often forget to check the simple problems before launching into a long and frus-
trating debugging session. For example, I found that capitalization is a major problem. Not
every platform is case sensitive, but enough are that you’ll want to check for this problem
before you begin troubleshooting something more complicated. If I suspect that any part of
my application will have problems with case sensitivity problems, I’ll include a routine that
checks for this problem before creating the SOAP server or instantiating other objects.
Security is also a major showstopper with SOAP. Many developers forget to include the
anonymous user, IUSR_<Server Name>, in the list of entities that can access the object. Of
course, you won’t have this problem if you disallow anonymous access. The point is that you
must include code that will report security problems accurately so the network administrator
doesn’t spend hours looking for a problem that he or she could fix in just seconds.
Some SOAP errors are toolkit specific. The WSDL files used in the example program con-
tain location sensitive information. If you move the WSDL file on the server, you’ll find
that your application no longer works. However, the error you get back indicates that there’s
an internal server error. Finding this problem can become frustrating because the WSDL
files contain a lot of information for even a small component. It normally pays to generate a
new WSDL file, rather than troubleshoot the existing one, if you suspect that someone
moved the file to another location on the server.
One of the odd errors that you may run into is compatibility problems between versions
of XML. For example, quite a few people have complained of applications that stopped
working suddenly when they installed Visual Studio.NET on their workstations. It turns
out that Visual Studio.NET upgrades the XML version installed on most machines. One
of the errors that this new version of XML looks for is the position of the XML tag
within a document. Previous versions of XML would allow the tag to appear in the mid-
dle of the document. This new version requires that the XML tag appear on the first
line. Of course, this is just one of many new errors that the new version of XML will
check. If you’re having problems getting a SOAP message transferred across the wire
after a tool update, it pays to check for these XML formatting errors. In many cases,
you’ll get an error message that has nothing to do with the actual problem, making it
difficult to find and fix.
 Performance Concerns for all Applications 117

Speaking of XML errors—make sure you check for attribute errors as well. For example,
many people have run afoul of the encoding attribute in the XML tag. Using UTF-8 encod-
ing with a machine that doesn’t provide the proper support results in a completely unread-
able message. Likewise, you need UTF-8 support when working with certain
languages—the old Industry Standards Organization (ISO) standbys (such as ISO-8859-1)
won’t work any longer for applications with a world view.

Reading the SOAP Specification

You’ll undoubtedly spend some time reading the SOAP specification during the development process and again
during the testing process. This book attempts to bring some clarity to some of the muddled sections of the
specification. It’s not a replacement for the specification, but it will help you understand the intent of some areas.
One of the more confusing areas of the specification is the use of the SOAP Fault element. We talked about the
composition of this element in the “SOAP Fault Messages” section of Chapter 2 and will talk about it again
here. This particular element has caused more than a little discussion on the various newsgroups that I fre-
quent because faults can appear in so many areas. Reporting an error properly requires knowledge of how the
SOAP Fault element works.
The SOAP Fault element can only appear once within the body of a SOAP message. You can’t create more than one
fault entry and the fault entry normally appears by itself without any other SOAP message data. Some newer SOAP
developers create one SOAP Fault element for each error, causing the client to reject the error message and leaving
the user in the dark as to the cause of a problem. Figure 2.3 shows a perfect example of a SOAP Fault element.
The specification also places limits on which sub-elements you can use to report certain types of messages.
This is one place where even veteran developers make mistakes. You always include a faultcode and faultstring
element as part of the SOAP Fault element. However, developers only use the detail element when the error Ch
appears within the body of the message. If the error appears within the header of a message, you use only the
faultcode and faultstring elements.
4
An even more important consideration about the detail element is that the server often uses it to report com-
ponent errors. If you look at Example 10 in the SOAP specification, you’ll notice that the faultcode and fault-
string elements report a general server error. The detail element reports the specific component failure that
caused the server error. Since the detail element is used for a specific component in this case, you’ll notice that
the namespace points to that component.
So, where do you place details about a SOAP header fault? They should appear in the SOAP header, if you
want to include any at all. The SOAP specification doesn’t require SOAP header error details and any imple-
mentation is up to the sender of the fault information. Unfortunately, the lack of a specific description of
header fault errors means that everyone is free to interpret them as they see fit. I think you’ll find that each
toolkit uses a different method for reporting header errors. Hopefully, the SOAP committee will fix this over-
sight in future specifications.
The point of this exercise is clear. Make sure you look at the examples if you use the specification to figure out
a design problem in your application. In some cases, the write-up isn’t enough to tell you precisely what you
need to know to solve a problem. This isn’t an unexpected problem since the SOAP specification is new. Many
of these areas left to the interpretation of the reader will disappear as the SOAP specification matures.

Performance Concerns for all Applications

Performance is troublesome when working with SOAP. As mentioned in Chapter 2, SOAP
has some performance inhibiting problems right now that I doubt a standards committee
will fix anytime soon because fixing them would add to SOAP’s complexity. Advanced
features such as connection and resource pooling are only available where the complexity of
118 Chapter 4 Using SOAP to Create a Simple Application

the underlying protocol allows for such features. This is counter to one of the design goals
of SOAP, which is to reduce the complexity of the programming environment for those situ-
ations where a complex protocol isn’t required.
SOAP applications respond to the same types of code optimization that all applications do.
For example, you want to perform as much work as you can outside a loop because perform-
ing work within a loop is expensive. Some optimizing compilers will even look for this com-
mon programming flaw and fix it within the executable for you. However, the fact remains
that you must perform some types of optimization yourself—an optimizing compiler just
can’t perform the required analysis.
In addition to the common code fixes, SOAP also responds to some fixes that reflect the
method used to create a SOAP message, decode it, or act upon its contents. That’s where
this section of the chapter comes into play. The following sections provide some handy tips
on how you can make your SOAP applications run faster. Not every tip will work in every
situation. You’ll want to test the technique with your application, see how it reacts, and then
use it if applicable. I’ve used simple demonstration examples for the sake of clarity.
Obviously, your code will contain complex interactions that you’ll want to check thoroughly.
Getting all of the fat you can out of your application not only makes the application run
faster, but could make it more reliable as well.

Attributes Versus Elements

One of the concerns that many developers have is the issue of using attributes versus
elements in their code. I’ve been using elements for all of the examples in the book
because they’re easier to read. In fact, you should use the element approach whenever
possible in your code for the same reason—to make the code easier to read. Here’s an
example of a simple SOAP message using the attribute approach. Note that the
MyObj:GetPerson tag attributes normally appear on one line, rather than multiple lines
as shown.
<SOAP-ENV:Envelope
xmlns:SOAP-ENV=”http://schemas.xmlsoap.org/soap/envelope/”
SOAP-ENV:encodingStyle=”http://schemas.xmlsoap.org/soap/encoding/”>
<SOAP-ENV:Header>
<t:Transaction
xmlns:t=”http://www.mycompany.com/myobjects/”
SOAP-ENV:mustUnderstand=”0”>
1
</t:Transaction>
</SOAP-ENV:Header>
<SOAP-ENV:Body>
<MyObj:GetPerson
xmlns:MyObj=”http://www.mycompany.com/myobjects/”
LastName=”Mueller”
FirstName=”John” />
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>

If you check this example at the SOAP Message Validation site at http://www.soaptoolkit.
com/soapvalidator/, you’ll find that it validates just fine. This model also works in real life.
The advantage of using the attribute approach is performance. It takes less time to parse this
 Performance Concerns for all Applications 119

kind of message than messages that rely on the element approach, which means you see a
small performance boost on a per client basis and a larger increase in overall server perfor-
mance.
As previously mentioned, one down side of the attribute approach is readability. Some
developers also feel that the attribute approach is less flexible. For example, you can’t pro-
vide type information when using the attribute approach. We discussed one problem with a
lack of type attributes in Chapter 2—you may find that servers like the Apache server won’t
parse the message properly.

Code Optimization
Performance is a function of the code you use to create the application. Sure, you can rely
on an optimizing compiler to do some of the work for you. For example, if you need short-
circuiting of the comparisons in an if statement, then an optimizing compiler is the tool of
choice. The developer can’t really optimize the code, in this case, because there isn’t any
way to know if short-circuiting will work until runtime. The compiler needs to insert the
short-circuiting logic.
However, there are times where the coding technique you use determines how well your
code will run. Consider the situation where you have to make a decision using structures or
simply use smart coding techniques to allow the application to make the decision automati-
cally. Developers use three common techniques to handle a situation where you have to take
a different action based on the input type of a variable. Ch

■ Create a series If...Then statements that test for the type of the variable and take appro- 4
priate action.
■ Use a Select Case statement to perform the comparison and perform the action within a
case.
■ Create a class containing the code and use overloading to perform the comparison auto-
matically.

All of these techniques work, but they’re based in part on procedural thinking. Sometimes,
all you really need to do is perform a type cast or use an existing method to perform the
work you need to do. Here’s a very simple example of using smart coding to reduce the
amount of time required to make a decision.
Private Sub Command1_Click()
Dim iTest As Integer
Dim sTest As String
Dim bTest As Boolean

CheckType (iTest)
CheckType (sTest)
CheckType (bTest)
End Sub

Public Sub CheckType(vType As Variant)

MsgBox “The input type is: “ + TypeName(vType)
End Sub
120 Chapter 4 Using SOAP to Create a Simple Application

I could have used the If...Then statement, a Select Case statement, or class to perform this
task. However, in this case, using something simpler is easier to code, easier to under-
stand, and works faster as well. This is an example of a programmer specific optimization
that you can’t perform using other techniques—an optimizing compiler won’t help you in
this situation.
The reason these techniques are so important in SOAP is that SOAP already has several
performance hits against it. The method you use to serialize data, encode and decode mes-
sages, and interact with it on the server all affect the performance of your application to a
great degree. Small changes to coding style can make big performance differences when it
comes to working with SOAP.

Project
Now that you have a better idea of how a simple SOAP application works, it’s time to put
some of that knowledge into practice. We began this chapter with a simple component that
could add two numbers together. However, it might be useful to try a component that can
do more. The following steps will provide an overview of an add-on to this project. Going
through these extra steps help reinforce what you’ve learned (or at least point out places
where you need additional study).

Depending on how you set up your server, you may need to shut down the Web ser-
vice and restart it to ensure the component you worked with earlier is unloaded. In
some rare cases, developers have reported the need to restart their machine in order
to clear memory completely. You’ll also want to clear the temporary Internet files from
your test browser so that you don’t receive false test indications.

1. Open the AddIt project.

2. Add methods for multiplying, dividing, and subtracting two numbers based on the code
we used for adding two numbers.
3. Copy the new version of the component to your test server and register the compo-
nent again.
4. Update your WSDL files using one of the WSDL generator programs we talked
about earlier.
5. Modify the client application to allow for multiplication, division, and subtraction.
6. Recompile the client and test the new capabilities of your application.

Notice that we didn’t modify the ASP file. Normally you won’t need to perform this step
when upgrading the capabilities of an existing application since the ASP file only acts as a
messenger. The WSDL and WSML files always require an update whenever you make any
kind of change to your server.
 CHAPTER
5
Migrating an Application from
DCOM to SOAP
In this chapter
Introduction 122
SOAP Application Conversion Prerequisites 123
Updating a Simple Utility Program 140
Updating a Data Viewer 145
Updating a Complete Database Application 153
Modified Application Concerns 155
Troubleshooting 159
122 Chapter 5 Migrating an Application from DCOM to SOAP

Introduction
For years, Microsoft has tried to convince you that COM, DCOM, and COM+ were the
grail of application development—that they could answer your every need. The reality for
most of us is that these technologies work great on a LAN, somewhat on a WAN, and
barely or not at all on the Internet. So, at this point, you may consider doing something rash
when the boss asks you to scale your application to work in a distributed environment. You
contemplate the huge investment you made in Microsoft technology and realize that it
won’t work in the new application environment. Don’t worry, all is not lost; you can still
recover. Converting all or part of your application to SOAP may be the solution you were
looking for all along.
I’ll warn you in advance that this chapter isn’t designed to provide you with step-by-step
instructions for converting a specific application to SOAP. There isn’t any way that I can
guess much about your application and the intricacies of converting it to a new program-
ming technology. However, this chapter will show you the basics of converting some appli-
cation examples that I feel represent basic application types, and you can use the knowledge
you gain to convert your particular application.
The first section of the chapter, “SOAP Application Conversion Prerequisites,” helps you to
understand what you need to consider before you begin converting an application. We’ll talk
about such important issues as determining which modules to convert. You probably won’t
want to convert the entire application, except in the most severe cases. Since you’ll be using
more than one protocol in the new application, you’ll need to consider methods for avoiding
protocol-related problems in the modified applications as well. There are also integrating
and binding issues to consider. This section talks about all of these issues and more.
The second (“Updating a Simple Utility Program”), third (“Updating a Data Viewer”), and
fourth (“Updating a Complete Database Application”) sections of the chapter show three
conversion scenarios of increasing complexity. I chose these three scenarios because they
best represent the types of applications you’ll likely need to convert. I do build on each
application as I convert it, so you’ll probably want to read all three sections.
The fifth section of the chapter, “Modified Application Concerns,” tells you about the three
major issues in a modified application. First, you need to consider reliability. The modified
application is using new code, so you need to test it as if it were a new application. Second,
since SOAP doesn’t provide the robust security of binary protocols like DCOM, you’ll need
to consider security issues for the new application. You may find that security needs require
you to split some modules into two parts: those that work with outside companies and those
that you can only use internally. Finally, the new application won’t run as quickly as before.
You need to decide if performance has become an issue and what you intend to do about the
performance problems. In some cases, you may decide to forgo using SOAP because perfor-
mance is too poor. If that occurs, then you need to work with partners to come up with
alternatives to the problem.
The final section, “Troubleshooting,” will show you how to troubleshoot converted applica-
tions based on what you’ve learned throughout the rest of the chapter. This is an especially
 SOAP Application Conversion Prerequisites 123

important section because diagnosing problems in a hybrid application (one that uses both
SOAP and a binary protocol) can be especially difficult.

Many developers who normally work with DCOM will need to make a quick transition
to SOAP as companies make the move to distributed application development. Not
every developer will need a novice-level book or a college course to start the learning
process. In some cases, an online tutorial will provide all you need. Many Web sites
provide these courses. For example, you’ll find a course, “XML messaging with SOAP”
at developerWorks (http://www-106.ibm.com/developerworks/). Just click the
Tutorial link and follow the instructions to select tutorial by name. This site does
require you to register before you can use it, so it’s likely that you’ll receive additional
e-mail from them.

SOAP Application Conversion Prerequisites

Ch
Converting any application from one technology to another is never a small undertaking.
Even simple application conversions require some level of advanced planning. Yet, some 5
companies will try to perform the conversion ad hoc and usually end up botching the entire
process. This section will help you avoid the fate of those who add SOAP to their existing
applications without forethought or planning. Each of the following topics will take you one
step closer to getting that application moved from the local server to a distributed environ-
ment where everyone benefits. (Of course, we’ll take the hands-on approach to the problem
later in the chapter.)

Lest you think that application conversion is only for those who have nothing to risk,
some major companies are already planning on conversions as I write this. One of the
more interesting companies is eBay. You can read a detailed report about its SOAP appli-
cation upgrade effort at http://scriptingnews.userland.com/ backissues/
2001/03/12#microsoftAndEbay. This article specifically states that eBay will use
SOAP in its upgrade efforts. Imagine being able to add eBay engine support to your next
application. Sometimes the future of development boggles the mind.

An Overview of the Conversion Process

It always helps to have a map when you begin a journey, and anyone who has converted
an application before knows that the journey is both long and filled with trouble. Heading
off into the great unknown without a map may have the appeal of becoming like the brave
adventurers of the past, but remember that many of those adventurers set out, never to
return home again. Unless you want to bear the scars of failed implementations, be sure
you have a complete map to guide you through the process.
So, where do you get a map before you begin your application conversion adventure? A
SOAP application conversion map consists of several elements. The list of the elements that
I feel are most crucial for success follows:
124 Chapter 5 Migrating an Application from DCOM to SOAP

■ A block diagram of your application.

■ A list of application conversion goals and requirements.
■ Schemas of any databases.
■ Descriptions of the interfaces between various application components.
■ Documentation of any security requirements for the system.
■ A list of end-user needs.

It’s usually a good idea to begin with the block diagram of your application. Outline areas of
the application that users outside the company must access and those that only users within
the company will need. For example, the code used to display data will probably require
both internal and external access. On the other hand, you may not need to provide any
administrative features outside the company, so these modules require internal access only.
In some cases, you’ll find a component that provides both external- and internal-only uses.
For example, some developers place the code required to display records in the same com-
ponent as the code required create new records. You can normally gain a slight performance
benefit, increase security, and create a modular application by splitting components along
external- and internal-only lines.
At this point, update your application’s block diagram. Add a list of interfaces between all
components. You’ll need this information later as you work with SOAP support files, such as
the Web Service Description Language (WSDL) files required for server-side components.
Some developers will use this opportunity to create WSDL or other IDL files as part of
preparing for the application upgrade.
➔ For details on using this particular technique, see the “Using Schemas as Design Tools” sidebar in Chapter 6.
It’s time to look at the interfaces—specifically at the data types used for method calls. You
may find that some data types don’t translate well to the world of SOAP programming. The
more complex the data type, the harder the conversion. In addition, some data types are
operating platform specific. You wouldn’t want to try to pass a window handle to an Apache
server that wouldn’t have any idea of what to do with it.
While the SOAP specification provides well-documented methods to handle variables such as
strings and integers, each vendor normally handles complex data types such as objects in
unique ways. Simplifying method call data types whenever possible now will make the appli-
cation conversion process easier and could net some performance gains. Of course, you have
to weigh the advantages of simplification against the time required to perform the conversion.
In addition, you only need to perform this task on data used externally—internal only com-
ponents won’t require any change unless you plan to use SOAP for the entire application.

The SOAP specification does include provisions for arrays, enumeration, and structures.
You can use these three complex data types to your advantage during the conversion
process. In many cases, you can group a set of parameters in a structure in such a way
that the structure elements describe the complex data type. This allows the remote
 SOAP Application Conversion Prerequisites 125

machine to reconstruct the complex data type in a form that it understands. In short,
SOAP provides ways around complex data types that allow you to maintain cross-
platform compatibility.

Depending on how you plan to use the application after conversion, you’ll probably want to
perform all of the required component upgrades at this point. For example, you’ll want to
separate external-use code into individual components by function and type. As you’ll learn
later, the smaller you can make the individual modules, the easier it is to apply individual
security and make changes later.
Compile the application and test it on a LAN using DCOM. Putting the application back
together and testing the new configuration now means that you’ll have one set of debugging
tasks out of the way. Using this technique also ensures you can check the new application
architecture using a development technique that you’re completely comfortable with, rather
than debugging the components using the SOAP.
Ch
Once you’re certain that the new application design works, that the application’s block diagram
is up-to-date, that all of the interfaces are set up properly, and that you’ve clearly delineated
5
external operations, it’s time to start the SOAP conversion. This may mean creating addi-
tional components. Some businesses use one component for internal use and another for
external use. For example, the external component might use a different processing technique
or include additional security. However, the majority of the conversion effort will include
creating listeners for the server and new clients or client-side components.
If you’re using the technique we talked about in Chapter 4, “Using SOAP to Create a
Simple Application,” to create your application, you’ll want to get all of your listeners set up
and ready for use. Many developers choose ASP for the listener files because they’re easy to
create and maintain. However, if you need the highest possible application speed and lowest
resource use, then going the ISAPI route is probably best.
The technique in Chapter 4 is great if you plan to write an application from scratch, but it
isn’t necessarily the best choice for a conversion. I find that the easiest method for converting
DCOM applications is to rely on the Microsoft SOAP Messaging Object (SMO) because it
allows you to simulate the existing environment within the client. The client and server
component code remain about the same using this method—the major change occurs with
the listener and SMO code you need to write. Whichever route you choose, be sure you
create your listeners, then create the WSDL files you’ll need to use with the application
➔ For details, see the “Microsoft SOAP Messaging Object Generator” section of Chapter 4.
One of the final tasks is creating or modifying the client. If you go the SMO route, you might
need to change the component instantiation method, but that’s about it. When using the
high-level or low-level API routes, you’ll need to write a new client and add existing business
logic to it. You can use any supported method for client access. Many developers will use the
client-side application approach we used in Chapter 4. However, you can also use Web pages
and rely on a scripting language to provide the required SOAP calls. The technique you
126 Chapter 5 Migrating an Application from DCOM to SOAP

choose depends on client needs, user requirements, and the number of client objects at your
disposal. The SOAP toolkit you choose makes a great deal of difference when it comes to
client-side access. You’ll learn more about this as the book progresses, especially in Chapter
10 when we begin to work with smaller devices such as Personal Digital Assistants (PDAs).

A Question of Globally Unique Identifiers

Anyone who has worked with any form of COM understands the concept of the globally unique identifier
(GUID). COM uses the GUID to keep track of various incarnations of components. Two components with the
same name won’t interfere with each other because both components have a GUID that COM uses to differen-
tiate them. Windows supports the GUID mechanism because it guarantees unique references.
SOAP also works with components and needs some way to identify them. However, there’s a difference
between SOAP and COM when it comes to using GUIDs. SOAP relies on a uniform resource identifier (URI) to
locate components. The URI points to the resource, allowing the application to find it as needed.
The problem is that URIs aren’t necessarily unique, which means you could run into component-naming prob-
lems. The general specification for URIs, RFC2396 (http://www.faqs.org/rfcs/rfc2396.html), states
there are actually two types of URIs. The type that most people are familiar with is the uniform resource loca-
tor (URL). URLs allow you to find a Web site, but could also be used to find resources on a local area network
(LAN). The second type of URI is the uniform resource name (URN).
Of the two types of URI, only the URN is a managed resource. Being a managed resource means that a URN is
guaranteed to be unique. Companies that want to use a URN must apply for a Namespace Identifier (NID)
from an authority such as the Internet Assigned Numbers Authority (IANA). The NID appears as part of every
resource reference that the company creates. You can read more about the format of URNs at http://www.
faqs.org/rfcs/rfc2141.html. RFC2611 (http://www.faqs.org/rfcs/ rfc2611.html) explains
how URNs guarantee unique references.
Does the non-unique nature of URLs mean that you should stop using them and switch to URNs? Developers
will continue to use URLs because they’re more flexible than URNs and you don’t have to go through a lot of
extra paperwork to use one. However, developers should try to create unique references by using URLs they
actually own.
Several groups are working on solutions to the problem of unique identifiers. One such group is Resource
Directory Description Language (RDDL). You can read more about this group at http://www.openhealth.
org/RDDL/. The RDDL specification shows how an organization could use an URL that it owns to point to an
XHTML document that contains a list of resources the company wants to make accessible. The document con-
tains a description of the resource in human-readable form and embeds the required machine information as
part of the description.

Deciding Which Application Modules to Change

Let’s look more closely at the problem of deciding which application modules to change.
The first factor to consider is which application modules to change based on external- versus
internal-only use of application features. The second factor to consider is that you might
need to change some modules to accommodate special needs such as additional security. Of
course, those are decisions that you’re required to make, but understanding that you need to
make them doesn’t really define the problem fully. It’s important know precisely which mod-
ules to change in the application and why you would need to change them before you begin
the conversion process. After all, you don’t want to make additional work for yourself.
Modifying a module that doesn’t require it can open security holes or reduce performance.
 SOAP Application Conversion Prerequisites 127

The following sections will help you understand the process of choosing the application
modules to change during conversion. Most companies find that this is the hardest part of
the process because it’s also the time where you have to develop a conversion plan. A bad
decision here will incur significant costs later in the development process.

Sometimes it’s difficult to keep up with all of the acronyms appearing on the scene
today, especially with something as fast-paced as XML. Acronym Finder (http://www.
acronymfinder.com/) does a good job of helping you decipher many computer-
related acronyms. It also pays to check out the XML Technology Protocol Reference at
http://www.xml.com/pub/a/2000/11/01/protocols/quickref.html.
Another place to find out what all of these acronyms mean is the WebServices
Resource Center at http://soap-wrc.com/ webservices/.

A Simple Scenario
Ch
Let’s look at a hypothetical situation where conversion of an existing application makes
sense. You manage a performing arts center. Business is OK, but it’s not great. The board of 5
directors and your investors would like to see higher attendance levels. Advertising might
help, but that still doesn’t make it easy for someone to get tickets, especially if they plan to
visit your city on vacation. In addition, advertising across the entire country is expensive and
there’s no good way to target your audience.
Exposing the list of performances and the mechanism for making reservations will help the
situation. Travel agencies could include a link to the services you provide to all of their
agents with little programming. A travel agent could look at the list of performances and
make a reservation for someone going on vacation with little effort. There are two advan-
tages in using this technique. First, every time you make things easier for people, sales will
almost certainly increase. Second, getting information to the people who can use it best will
also result in increased sales at a low advertising cost.
Of course, you wouldn’t want to expose anything other than the list of performances and the
ability to make reservations. You might not want to provide the capability to cancel a reser-
vation without a call, for example. That’s why it’s important to decide which parts of your
applications to expose using SOAP. You only want to expose the functionality that other
people can use, not something that could potentially harm your business. Even if a third
party might need to cancel a reservation, forcing them to do so via a telephone call could
reduce the potential for cracker activity on your system. In addition, using this method
would tend to discourage cancellations.

Differences in Viewpoint
Another part of the selection process is choosing modules for the right reason. Distributed
application development requires a change in perspective. Squinting can help, but under-
standing how distributed applications differ from the desktop or client/server variety you’re
used to working with is better.
128 Chapter 5 Migrating an Application from DCOM to SOAP

First, you need to consider that you’re no longer relying on a two-tier or three-tier pro-
gramming model. The Internet isn’t a peer-to-peer programming environment either
because an application must store data in a centralized location. Developers refer to the
Internet as an n-tier model because clients do some of the processing, while servers handle
other parts. A server might need to call upon other servers to handle a request. SOAP is all
about exposing services to the Internet community in whole or part. You’ll receive requests
for service from a variety of sources, making it difficult to create a picture of activity outside
your organization. It’s impossible to make assumptions in this environment.
This brings us to the second change in perspective. Huge, monolithic components won’t do
the job in the Internet environment. SOAP and other Internet technologies work better
when a component has a single well-defined purpose. For example, some companies will
place all of their common dialogs in a single component file when working with desktop
applications. You would need to using individual, single dialog, files on the Internet. If you
can create a component that performs one task well, it becomes easier to create distributed
applications. Unlike a desktop application, a distributed application should only load the
functionality required to perform a task.
A third perspective-changing issue is that SOAP is stateless. You need to create complete pack-
ages. Making multiple calls as some DCOM applications do won’t work in the SOAP environ-
ment. It’s hard to save state implementation and doing so will lead to significant problems.
A fourth problem is trust. Other people will rely on your application if you expose it as a Web
service. This means that the components you provide can’t change. Any change in interface will
break other people’s code. The application also has to provide 24/7 accessibility. Depending on
how many people use your application, you’ll likely find that people are hitting your Web site at
all hours of the day and night.

Dividing the Application

When you look at your application, some external use needs become obvious early. For
example, if you’re converting an order entry system to allow customers to check on the status
of their order, then you’ll definitely want to convert the module that allows someone to dis-
play an order. However, you don’t want the customer to have ability to change the order in
any way, so you wouldn’t make the module that allows order entry accessible externally.
The order entry scenario does raise an interesting question. Although you don’t want the
customer to change an order using the same methods as employees, you may want to give
the customer the choice of canceling orders. This is one place where creating a smaller
component that performs one task well comes into play. A client/server developer would
place the ability to cancel orders within a large component that contained other features,
such as the ability to create new orders. Distributed application development would require
you to create several small components to do the work.
This order cancellation component will react differently depending on how a client accesses
it. Instead of allowing direct access to the main database when working with an external
client, you’ll want to create a customer cancellation database that you can check for cancel-
lations later. This allows you to validate the cancellation with the customer before making it
 SOAP Application Conversion Prerequisites 129

permanent in the main database. Just remember the principle of creating small components
that perform a single task exceptionally well, and you’ll find that distributed application
development is easier.
Eventually, you’ll end up with a list of coding projects based on the decisions you’ve made about
internal and external component access. You can easily divide these projects into four areas:
■ Unmodified components that you won’t change because they work fine as is and
no one will access them externally. Be certain you don’t have to duplicate the func-
tionality of these older monolithic components anywhere else in your application. If
you do, consider splitting the component into functional parts and reworking it.
■ Modularized components that contain few changes in business or procedural
logic. For the most part, you’re splitting a large component into its constituent parts so
that you can keep functions required by external processes separate from those that only
internal processes will use.
■ Specialized components that react differently based on client. These components
require more work to modify. The differences in reaction support security, privacy, or Ch
other system requirements. 5
■ New components that support either DCOM or SOAP services in a unique man-
ner. You’ll find that the vast majority of new components support SOAP functionality.
In many cases, these new components provide client-side processing support. This is by
far the hardest component programming for the application since the component will
need to perform SOAP specific tasks. Low-level API components are much harder to
develop than those that rely on SMO or the high-level API.

After you have a project list put together and have the tasks categorized by difficulty, you’ll
want to start assigning them to members of your development team. The fact that each
component and the interfaces it supports are well defined should allow team members to
work in tandem. Of course, you’ll want to perform several levels of integration testing as
team members complete their assigned tasks. Creating a critical time line that shows which
components to develop first will make it easier to perform integration testing as needed.

Avoiding Protocol-Related Problems in Modified Applications

The application you created using DCOM relies on the features provided by that particular
protocol. Remember that DCOM provides strong security, the ability to use any data type
within the application, and tight integration with the Windows operating system. What this
means to you as a developer is that you might run into situations where the assumptions
made in the original application conflict the realities of working with SOAP. Unlike
DCOM, SOAP doesn’t provide complete access to every data type, comes up short in the
security department, and lacks much of a connection with the underlying operating system
at all. SOAP also works in a distributed application environment, which means your code
might appear in public view. In short, converting an application is likely to cause some
problems on both the client and server end.
130 Chapter 5 Migrating an Application from DCOM to SOAP

One of the biggest problems that you’ll run across when working with distributed applica-
tions in an environment that isn’t entirely under your control is the use of other languages.
The problem is that your application might have to work with code from other countries.
The SOAP specification is unclear on the ramifications of foreign-language use and the
encoding requirements for your code. However, RFC2376 (http://www.normos.org/ietf/
rfc/rfc2376.txt) and the newer RFC3023 (http://www.normos.org/ietf/rfc/rfc3023.txt)
both provide specific encoding requirements. You need to decide what type of encoding to
support at the outset of your project. Not including the proper encoding will result in an
application that fails unexpectedly when used with other languages.

As of this writing, the Apache toolkit is the best one to use for international applica-
tions. It provides the broadest range of international support, including the use of
proper encoding in all generated files. The support isn’t perfect, but it’s far better than
any other SOAP toolkit available at this time. Since language support is such a hot
issue, expect to see vendors update their SOAP toolkits in the near future. Microsoft, in
particular, is anxious to update its language support. However, it’s important to verify
how the vendors update that support and whether it follows the proper specifications.
For example, following RFC3023 is better than following RFC2376 alone. You can still
hand edit the output of any toolkit—that’s the beauty of using a text-based protocol.

Of course, language support is just one of many problems you’ll face. The following sections
will help you decipher protocol-related problems in your converted application and will tell
you how to avoid them. Of course, not every problem is the simple protocol error that it
first appears to be. In many cases, you’ll find an underlying problem that wasn’t obvious at
first or that only appears in certain circumstances because of branching in your component.
In fact, odd errors that only occur when a given set of conditions is present are one of the
major concerns for developers. The SOAP protocol is relatively easy to understand, but
finding those odd errors that result from branching conditions can prove troublesome.

We’re examining a limited number of SOAP implementations in this book mainly

because I would need several books to discuss them all. You might find that none of
these SOAP implementations precisely meet your needs and that you would prefer to
use something else. Finding these other SOAP implementations could prove difficult,
even with the extensive search capabilities of the Internet. Fortunately, someone has
already done part of the legwork for you by creating a list of current SOAP implemen-
tations at http://www.soapware.org/ directory/4/implementations.

Server-Side Component Assumptions

Not every protocol problem you run across will resolve itself to something within the SOAP
message, runtime files, or associated components. Some developers have noted odd problems
with their SOAP implementations. For example, one developer recently noted that when
using the high-level API calculator example in the Microsoft SOAP Toolkit, he would get an
 SOAP Application Conversion Prerequisites 131

“Object Required” error message. Yet, if he commented the error parsing code out of the
listener, the example would work. He then noted that he had made a few small changes to
the example for the sake of experimentation.
The problem turned out to be one where the error occurred within the component after it
had already generated the correct response. Although the problem appeared to be within the
SOAP portion of the application, it actually resided in the component in a location that wasn’t
immediately apparent from the error message provided. The problems with all of the connec-
tivity that SOAP requires is that an error can hide. You might think the error is in one loca-
tion when it actually appears in another. For this reason, you’ll want to know how to test your
application using several techniques to ensure you can view problems from several perspec-
tives. That’s one of the reasons why I’ve provided so many test Web sites in previous chapters.
The developer in this scenario could have avoided the problem by testing the component
locally before deploying it on the server. He could have also tried using the component with
other protocols to see if the problem actually occurred within the protocol or as part of the
application component. Using these two tests would have avoided many hours of trou- Ch
bleshooting time because the developer would have seen the error locally. In short, the best
way to avoid some protocol-related problems is to perform proper module level testing. 5
Another problem is that SOAP is actually a message formatting specification and nothing more.
Developers can assume many things about the SOAP specification, most of which aren’t part of
the specification at all. For example, SOAP doesn’t provide transaction support and it won’t
provide transaction support anytime in the future. If you need transaction support for your
application, then you need to choose a transport protocol that supports transactions such as
DCOM or CORBA. The combination of protocols you choose is important because the server-
side component will rely on these protocols to deliver all of the required information. Even if
the application delivers a correct SOAP message, that message might not arrive at the compo-
nent if the application doesn’t use all the required protocols to deliver essential information.

Error Handling Code

Moving your error handling code to a special module designed for the process may seem
like overkill, but this is one area where a lot of developers run into trouble. DCOM provides
one set of error handling resources; SOAP provides another. You can’t use a single set of
error handling routines to handle both protocols so it’s best to place them in separate mod-
ules. Let’s look at a typical example of the differences between SOAP and DCOM when it
comes to error handling. Here is a short module designed to throw an error. Notice that I’m
not using anything odd in the way of code.
Public Function ThrowError(myInput As String) As String
‘Generate some quick output.
Dim myOutput As String
myOutput = “There was an error! Bad string: “ + myInput
ThrowError = myOutput

‘Create an error message.

Err.Clear
Err.Raise 100, App.Title + “.GenError.ThrowError”, “This is an error!”
End Function
132 Chapter 5 Migrating an Application from DCOM to SOAP

You’ll find all of the sample code in this chapter in the Chapter 5 folder on the Que
Web site for this book. You can find it at www.quepublishing.com. The examples
contain more code than shown in the book—the listings show only the code required
for explanation purposes.

You’ll notice that the example sets the function to a value, and then throws an error.
Theoretically, the only thing that we’ll see as output is the error information.
Here’s the code that I’m using on the client side. The error handling is a little more exten-
sive than before because I want to show you how much of the error information actually
arrives from the component. In other words, we need to answer the question whether SOAP
hides anything.
Private Sub cmdThrowError_Click()
‘Create the SOAP client.
Dim Client As SoapClient
Dim Result As String
Dim ErrorMessage As String

‘Set up an error handler.

On Error GoTo ErrorHandler

‘Create the connection.

Set Client = New SoapClient
Client.mssoapinit _
“http://winserver/soapexamples/ShowError/ErrorSim.WSDL”, _
“ErrorSim”, _
“GenErrorSoapPort”

‘Initiate the error.

Result = Client.ThrowError(“Hello World”)
MsgBox Result, vbOK, “Error Message”
Exit Sub

‘Display a message when an error occurs.

ErrorHandler:

ErrorMessage = “Fault Code: “ + Client.faultcode + vbCrLf + vbCrLf + _

“Fault String: “ + Client.faultstring + vbCrLf + vbCrLf + _
“Fault Actor: “ + Client.faultactor + vbCrLf + vbCrLf + _
“Detail: “ + Client.detail
MsgBox ErrorMessage, vbExclamation
End Sub

This code is similar to the client code I created in Chapter 4. The main difference is that
we’ll look at all the potential fault information. The application formats the output so that
you can best see the available information. Figure 5.1 shows the output from this example.
 SOAP Application Conversion Prerequisites 133

Figure 5.1
The fault information
output from this
application isn’t
complete.

Look at the error generation code in the component listing and compare it to the output
shown in Figure 5.1. You’ll notice that you didn’t receive the error number. If this had been
a DCOM application, you would have seen an error number that might provide more infor-
mation than the text. Figure 5.2 shows the data flow between client and server (I’m using Ch
the Microsoft SOAP Trace Utility in this case.)
5
Figure 5.2
The fault information
from a component is
much more detailed
than from a
SOAP error.

As you can see, there’s a lot of information in the message, but the error number doesn’t
match the one output by the application. The fact that SOAP outputs a lot more informa-
tion for a component related failure is great, but it doesn’t replace the information that’s
missing. This is one of the reasons you’ll want to use multiple error handling routines. A
SOAP application component will need to output more information as part of the descrip-
tion than your DCOM components will. Notice that the output does contain the name of
the component as the application name. That’s the effect of the App.Title entry in the
134 Chapter 5 Migrating an Application from DCOM to SOAP

component code. You should always provide this information so that you can track down
multilevel component failures with a little less trouble.
Of course, using multiple help routines raises another set of problems. You need to consider
how the component will know which error handling routine to call if you create a mixed pro-
tocol application. The technique that I rely on is to add another argument to the component.
This argument remains undefined when DCOM calls the component, but contains a simple
text value when called by a SOAP listener. In this way, all the new code resides with the SOAP
listener and the DCOM client application continues to operate essentially as it did in the past.

I’ve mentioned many times the SOAP specification is changing as I write this and will
probably continue to change for quite some time. At least some of these changes are
the result of errors or omissions within the current specification. In other words, don’t
always assume that error messages you see are problems in your code—they may be
failings of the specification. You can find the current SOAP specification problems at
http://www.w3.org/2000/xp/Group/xmlp-issues. You’ll also want to consider
interoperability issues between toolkit implementations of SOAP. Find out more at
http://www.soapware.org/ interopathonPlan.

Some modules require modification simply to accommodate differences in protocol imple-

mentation. We’ll see later that you should create separate error handling modules for a mixed
mode application. The requirements of SOAP error handling are quite different from those
needed for DCOM. Since an application developer normally assumes a single error handling
technique, you’ll find that you need to create these separate modules as new entities and find
some consistent method for accessing them. In short, you need to consider protocol require-
ments when determining which modules to modify. It’s important to ask the question of how
the new protocol, SOAP, acts differently than the old protocol.

Integrating New Modules with Existing Application Elements

The objective of any programming exercise (at least the eventual goal) is to create an end-
to-end coherent application that performs all the tasks the users request of it. In the case of
application modification, you begin with a fully functional application and end up with a
fully functional application that has new features. In between these two stages you have
various stages of dual application development. In other words, you’re actually creating
two applications: one that works with SOAP and another that works with DCOM.
So, how do you finally create one application from two? Developers take many approaches
to the problem. The method that seems to work best is to begin with the initial application
and add task-related functionality to it. We’ll see how this works later in the three example
applications. The concept is simple, but the implementation can be difficult.
Individual component testing can also help. Ensuring that each component is bug free before
trying to integrate it with the larger application is important, especially with SOAP because
 SOAP Application Conversion Prerequisites 135

there are many outside factors that can affect test results. To give you some idea of how bad
this problem can be, one developer I talked with spent the better part of two days trying to
diagnose a problem with an application. His thought was that since the component was so
simple, the problem had to be with the network setup or some other factor. Even looking at
the code didn’t seem to help, in part because he was so convinced the problem had to be
elsewhere. A local test, however, showed that the component was faulty. If the developer had
tested the component locally before integration, this problem wouldn’t have happened.
Testing the individual component with a reliable transport on a test LAN is also helpful. In
this way, you can eliminate network specific errors. A component that works locally may not
work across the network. I normally test my component with DCOM first, because I’m
familiar with errors that DCOM will present. I then test locally using SOAP. Remember
that you won’t fully test the component with DCOM because there are some SOAP-specific
issues to address in the code.
This may seem like a lot of additional work. Developers under a deadline (who isn’t today)
often skip these intermediate steps hoping the component will work the first time. More Ch
often than not, skipping the steps only wastes time because the developer ends up doing
them again anyway. 5
Integrating new modules into an existing application can be tricky for other reasons. The
most important consideration is production system downtime. You want to keep downtime
to a minimum. All of these intermediate steps allow you to test the component fully outside
the production environment. When you do add the new component to the system, the users
should see added functionality and nothing else.
Security is also an issue when adding new components to an existing application. You don’t
want everyone to try the component at the same time. The testing phase should include a
timeframe in which experienced users test the functionality provided by the new component
and provide comment on how well it integrates with the system as a whole. Often, a compo-
nent works completely from a programmer’s perspective, but breaks the application from the
user’s perspective because of difference in workflow. SOAP applications work in a distributed
environment, so make sure you test the application in every environment and use every
device. This is something that DCOM developers might not have worried about in the past
because a DCOM application usually resides on a desktop client.
This final piece of the puzzle, user perspective, is actually the most difficult part of the inte-
gration. Some developers try to handle it by telling users that this is the new way the appli-
cation works and they can’t do anything about it. One developer went so far as to restrict a
user’s ability to handle commands at the keyboard, rather than use a mouse, just to deliver
an application on time. The company didn’t accept the application because of user complaints
about the change in workflow.
Once the experienced test group completes their testing of the new functionality, open it up
to the company as a whole, at least those users who actually need to use the new application
features. This last phase of testing should go smoothly if you’ve performed all the required
intermediate steps. However, you should still solicit responses from the users, especially if
136 Chapter 5 Migrating an Application from DCOM to SOAP

your application will run on new types of clients such as PDAs. (We’ll look at a PDA exam-
ple in Chapter 10, “Working with PDAs.”) In some cases, you’ll find that you need to make
changes to the interface to accommodate these new device types—devices that you haven’t
had to worry about when working with DCOM.

Implementing SOAP with COM Language Binding

SOAP transfers data from one point to another. It’s nothing more than a messenger when
you think about it. The capabilities of SOAP are useless if the message is garbled or in a for-
mat that the recipient can’t understand. The process of binding allows the client to discover
the message format required by the server to exchange information. SOAP provides the
conduit required for this discovery phase.
Of course, there are a number of ways to provide binding services using SOAP. The SOAP
specification doesn’t say much itself and leaves this detail up to the implementers. Microsoft
and other companies choose to use WSDL as the discovery mechanism for SOAP. When
you write the client portion of an application (as we did in Chapter 4), the application first
queries the WSDL file to discover the message format required by the server. When you
follow the flow of information from client to server, you’ll discover that the server sends the
WSDL file to the client so that the client can use it for message formatting. (Make sure you
read the sidebar, “Is WSDL Sufficient?” to discover some additional thoughts about this
particular method of binding the component to the client application.)

Is WSDL Sufficient?
We looked at a simple SOAP example in Chapter 4 and used a Web Services Description Language (WSDL) file
to define the component interface for that example. Just about every component-based application today relies
on some type of Interface Definition Language (IDL) to describe the component’s construction. WSDL seems to
work fine in the Chapter 4 example, but the component used provides such a simple interface that we probably
didn’t need even WSDL to define it. However, what happens when you begin working with complex compo-
nents or need to convert applications as we’re doing in this chapter? Does WSDL provide a sufficiently robust
specification that it can handle these situations?
Some developers are asking these questions as they work with WSDL. The WSDL specification is relatively new.
Vendors designed the WSDL specification to answer the problem of working with multiple proprietary IDL spec-
ifications such as Service Description Language (SDL), Network-Accessible Service Specification Language
(NASSL), SOAP Contract Language (SCL), and SOAP Interface Definition Language (SIDL).
The short answer to the question of whether WSDL is sufficient is no. It won’t handle all the needs of a devel-
oper today and there’s a lot of room for interpretation in the specification. Because of these problems, some
developers are proposing yet more specifications, such as A Little Interface Definition Language (ALIDL). You
can read about this specification at http://www.xmlrpc.com/alidl. Of course, WSDL also has competition
in the form of the established Universal Description, Discovery, and Integration (UDDI) specification.
The long view of technology is different, however. A developer today needs to wade through vast numbers of
specifications because the technology is in its infancy. As SOAP and its associated technologies mature and
developers have a larger knowledge base with which to make decisions, the current problems with WSDL
should resolve themselves. In the end, developers need to choose a technology and use it until the technology
matures or a truly better technology comes along. WSDL seems to provide the best IDL available for SOAP
today—it’s the right choice for most development projects.
 SOAP Application Conversion Prerequisites 137

WSDL isn’t the only method for providing SOAP binding. Many developers prefer to leave
nothing to chance and address the issue as a design decision. The SOAP message itself will
always provide sufficient information for binding to take place. Of course, this requires an
intimate knowledge of the application as a whole and provides a static solution to the prob-
lem. You’ll see in Appendix A, “SOAP Data Types and Data Type Conversions,” that this
decision also affects the way data types are handled and can even change the performance
characteristics of the application.
Some products, such as 4S4C (see Appendix D, “SOAP for Visual C++ Developers,” for details)
directly query the component and provide binding in that manner. This is reminiscent of the
method used by DCOM and will be the most familiar to Visual C++ developers who have had
to worry about binding details in the past. Few vendors are embracing this methodology, but it’s
an important method to consider if you have a large, well-established, application.
No matter which method you use to bind the client to the server, SOAP will act as the mes-
senger. The method used to perform the binding will affect the SOAP message format, but
the result is the same—a client talking to a server and gaining access to resources. Vendors Ch
normally choose a binding method and don’t provide a means to modify that choice. As a
result, you need to consider the choice of toolkit wisely to avoid problems with the current 5
application setup.

Productivity Tips
If you’re in the same position that most DCOM developers are, the company president is
breathing down your neck asking every hour about the progress you’ve made in converting
the company applications for Internet use. It doesn’t matter that it took years to write those
applications—the president wants them converted today in order to keep up with the com-
petition. You need to increase productivity as quickly as possible.
This section of the chapter won’t include all of the productivity enhancements I’ve
researched. You’ll find them spread throughout the book in the appropriate places. The pur-
pose of this section is to look at some productivity aids for those people who are converting
applications. Combine your favorite tips in the following sections with those in the rest of
the book to come up with some techniques that work best for you.

Quick Research
Chapters 1, “An Overview of SOAP,” and 2, “SOAP in Theory,” provided you with some
ideas on how a typical SOAP message is put together. In Chapter 4 we looked at how appli-
cations create SOAP messages to transfer data from one place to another. However, you
might not have time to read all of that information every time you want to find out the
meaning behind a particular SOAP element. You might need the information now. That’s
where the SOAP 1.1 Reference Web site (http://www.zvon.org/xxl/soapReference/Output/
index.html) comes into play. This Web site divides the SOAP message into easily recogniz-
able pieces and allows you to read about that element only. Figure 5.3 shows what this Web
site looks like.
138 Chapter 5 Migrating an Application from DCOM to SOAP

Figure 5.3
The SOAP 1.1
Reference Web site is
the best place to find
information about
SOAP quickly.

As you can see, all you need to do is choose a SOAP message element in the left pane to see
information about it in the right. You can also click on keywords within the description,
such as the encodingStyle keyword shown in Figure 5.3. This Web site only provides a
quick look at SOAP elements, however. If you want to learn detailed SOAP information,
you still have to rely on the SOAP specification or detailed write-ups such as those found in
Chapters 1 and 2. Clicking the Go to standard link on this Web page takes you to the
description of that particular element in the standard. It will still require time to read about
the element, but at least you won’t spend time searching the specification for it.
While the SOAP 1.1 Reference Web site won’t save you astronomical amounts of time, it
does help in a world where every second counts. You can find the information you need,
written in a very terse style, extremely fast. This is the place to go when you’re in a hurry
and don’t care to know all of the details about a particular SOAP element.

Improving Your IDE

One of the best productivity enhancements you can perform is to improve the usability of
your IDE. Many developers make their special tools available online—free for the price of a
download. One of the better tools that I’ve found for Visual Basic is the MZ-Tools add-in
(http://www.mztools.com/). This particular tool enhances productivity by adding a Main
menu to the Visual Basic IDE, along with several context menus. We’ll explore this tool in
detail in Appendix C, “Third Party Tool Reference.”

Researching the WSDL Files

You’ll have to create WSDL files for all of your existing components and any new ones
required for SOAP use. Studying these WSDL files can often point out flaws in an imple-
mentation that you might not otherwise see. Let’s look at a specific example.
 SOAP Application Conversion Prerequisites 139

Figure 5.4 shows the WSDL file for the AddIt application in Chapter 4. Notice that the
application passes the two numbers used for the addition both to and from the server. While
this small amount of data won’t cause a big performance drop when using DCOM, it could
cause a significant problem when working with SOAP over an Internet connection.

Figure 5.4
The AddIt component
WSDL file shows an
implementation flaw.

Ch

All we really need for a return value is the result. The client already knows the two numbers
used to create the result and there isn’t any reason for the server to modify them. Here are
the simple changes you’d make to the application to fix this problem (the changes are in
bold type).
Public Function DoAdd(ByVal Add1 As Integer, _
ByVal Add2 As Integer) As Integer
‘Calculate the result.
DoAdd = Add1 + Add2
End Function

No, it’s not a very big change, but it can make a difference in the operation of your SOAP
application. Figure 5.5 shows the new WSDL file for this example. You may not see a very
big difference with this example, but it can make a huge difference when you have several
hundred users all hitting your database application at once. Even small performance gains
can make the difference between an application that users like and one that gets complaints.
The bottom line is that researching the WSDL files can help you determine that the
interfaces you think you have in place are the ones that the application will actually use.
Verification of the current state of your DCOM application is an important part of the con-
version process. Using this WSDL technique can save hours of looking at source code
because it provides you with an encapsulated view of your application in an easy-to under-
stand format. Theoretically, you could even build an application to parse the WSDL files
140 Chapter 5 Migrating an Application from DCOM to SOAP

and verify them against your application design, saving even more time. (I imagine a third-
party developer will eventually create an application just like the one I’m talking about here,
so keep your eyes open.)

Figure 5.5
We fixed the flaw with
a small coding change.

Updating a Simple Utility Program

Utility programs come in a number of shapes and sizes, and people invent new forms of
utility programs every day. Microsoft recently introduced yet another version of the utility
program, the Web Service. No matter which utility program you’re talking about, they all
seem to have several common elements.
■ Small size
■ Not a main application
■ Easy to understand and use interface, or no interface at all
■ Only one or two modules with few interfaces
■ Normally used by savvy users

This section of the chapter will look at a simple utility that you might want to convert from
DCOM to SOAP given today’s distributed environment. This utility simply polls the server
for the current time and date. (We’ll extend this with the capability to poll service status in
Chapter 6, “Creating Remote Access Utilities.”) Sure, you can get this information using
other methods, but having an agent on every server that sends the information back to a
small utility can be quite useful, especially if you have more than one type of server.
 Updating a Simple Utility Program 141

We talked about the need to use WSDL files with the Microsoft SOAP Toolkit in Chapter
4. The toolkit comes with two utilities you can use to create WSDL files. The problem
with these utilities is that they generate UTF-16 files that may not work in every situa-
tion. You can find another fully functional WSDL Generator at http://www.pha-
lanxsys.com/soap/wsdlwiz.htm. I’ll also discuss this utility in Appendix C.
Another good WSDL toolkit candidate is the one from alphaWorks that you’ll find at
http://www.alphaworks.ibm.com/tech/wsdltoolkit.

Updating the Server-Side Component

DCOM components start life as EXE files. That’s because Windows executes them as out-
of-process components. An out-of-process server doesn’t rely on anything else to work. All
you need to do is create a registry reference to it and then access it from within a client
application. The local registry reference points to the server that contains the components,
while the remote registry reference contains the normal out-of-process server entries.
Ch

The Que Web site contains all of the code used in this example. You can find it at
5
www.quepublishing.com. You’ll find the DCOM client and component that I used as
a starting point in the \Chapter 05\DCOM Utility Client and Chapter 05\DCOM Utility
Component folders. The converted client and server appear in the \Chapter 05\SOAP
Utility Client, \Chapter 05\SOAP Utility Local Component, and \Chapter 05\SOAP Utility
Component folders.

You can begin the conversion process for a Visual Basic project by opening the Project
Properties dialog box shown in Figure 5.6. Change the project type to ActiveX DLL and
check the Unattended Operation option. These two changes are all you need in many cases
to convert the project from an out-of-process to an in-process server. Visual C++ users will
want to create a new project and copy the class specific code. In either case, create the DLL
form of your component and move it to the server.

Figure 5.6
Modify the Project
Properties dialog to
match the new com-
ponent requirements.
142 Chapter 5 Migrating an Application from DCOM to SOAP

In some cases, the component won’t compile correctly, especially if it references other com-
ponents. Visual C++ developers will experience this problem more often than Visual Basic
developers. Remember that all references are from an in-process server now. This places
limits on what you can do with the component from a memory usage perspective since the
component won’t have its own memory to use. You also have to check threading issues. The
change from out-of-process to in-process means that you may be able to squeeze a little
extra performance out of the component by changing the threading model to match the
host. Theoretically, you won’t have to consider any additional thread-safety issues if the
component is already thread safe in its EXE format. However, some DCOM developers
take shortcuts with thread safety, so you may run into problems with this issue as well.
Of course, you’ll need to generate the requisite WSDL file for your server. I find that the
SOAP Toolkit 2.0 Wizard dialog shown in Figure 5.7 provides a good overview of problem-
atic methods. It highlights any methods with problem data types in red. If you see any prob-
lem methods, you may have to choose a different data type for your component method and
upgrade the client as well. These problem methods will also generate question marks,
“????????” in the WSDL file. If you don’t see any red methods, then generate a WSDL file
and check it for problem methods. I talked about this process in the “Researching the
WSDL Files” section of the chapter.

Figure 5.7
The SOAP Toolkit 2.0
Wizard dialog will tell
you about any prob-
lem methods in
your component.

Creating the Local Component

The client application is still expecting a direct component reference. The first method for
handling this problem is to change the client to use the WSDL file. You’d use code similar
to the code found in Chapter 4 to go this route. The second method is to use a local compo-
nent that accesses the remote component using SOAP.
 Updating a Simple Utility Program 143

You can use two different techniques to create the local component. The first method is
to use the SOAP Messaging Object (SMO) technique. I introduced the Microsoft SOAP
Messaging Object Generator in the “Microsoft SOAP Messaging Object Generator” section
of chapter 4. Don’t get the idea that this method is a free ride since the wizard creates a lot
of the code for you. It does greatly reduce the amount of work, but you’ll still need to per-
form a lot of coding. SMO is more suitable for database application development, so we’ll
look at it later in the book.
The second method is to create a local component that stands in for the remote component
that you used in the past. The local component will access the remote component using
SOAP. That’s the technique most suited to an existing DCOM application. If you give the
local component the same name and use the same interfaces as you did before, the client
application might not require much in the way of update. Best of all, except for a local regis-
tration that you could perform using a script when the user logs into the server, you don’t
have any messy updates to perform.
Begin by creating an ActiveX DLL or ActiveX EXE application. You may be able to gain a Ch
speed advantage using an ActiveX DLL, but the application update process is simpler using
an ActiveX EXE. I normally rely on an ActiveX EXE to make things simple. Here’s the code 5
you’ll add to the project.
Public Function GetTime() As String
‘Create the SOAP client.
Dim Client As SoapClient

‘Set up an error handler.

On Error GoTo ErrorHandler

‘Create the connection.

Set Client = New SoapClient

Client.mssoapinit _
“http://winserver/soapexamples/DCOMUpdate1/CheckStat1.WSDL”, _
“CheckStat1”, _
“CheckServerSoapPort”

‘Get the status information.

GetTime = Client.GetTime

Exit Function

‘Display a message when an error occurs.

ErrorHandler:

‘Create the error message.

Dim ErrorMessage As String
ErrorMessage = “Fault Code: “ + Client.faultcode + vbCrLf + vbCrLf + _
“Fault String: “ + Client.faultstring + vbCrLf + vbCrLf + _
“Fault Actor: “ + Client.faultactor

‘Pass the error back to the client.

Err.Raise 0, “SOAP Processing Error”, ErrorMessage
End Function
144 Chapter 5 Migrating an Application from DCOM to SOAP

Notice that the code is similar to what we used for the client application in Chapter 4. You
still have to create a SoapClient object and use it to provide the remote access. Obviously,
you can use any other technique supported by the Microsoft SOAP Toolkit as well, but this
method works well for DCOM upgrade projects where you don’t have a lot of data con-
cerns. Of course, instead of displaying the results within the component, you’ll send them
along to the client. Also, notice that this example provides more error handling information
to the client, but that the error number is still set to 0 because we can’t determine it from
the SOAP message.
The example uses the same project and class name as the DCOM project that I updated. In
fact, if you look in the DCOM Configuration Utility (Figure 5.8), you’ll see two copies of
the component now. This brings up an important point. Make sure your registration script
for the new local component also contains code to unregister the old remote component.
Otherwise, the user will end up with ambiguous registry entries and you might end up with
additional work in the form of support calls.

Figure 5.8
The DCOM
Configuration Utility
will show two ver-
sions of the same
component.

Updating the Standard Client

Given the method that we’ve used to create the local component, updating the client is easy.
All you need to do is change the reference within the Visual Basic application. I gave the local
component a distinctive name so that it would be hard to confuse it with the old remote com-
ponent. Figure 5.9 shows the one and only change to the updated DCOM client.
At this point, all you need to do is compile the client and distribute the new copy to users.
Theoretically, you could carefully code the local component so that you wouldn’t need to
change the client at all. What this means is ensuring the local component has the same
GUID as the old remote component did. You’ll also need to make the appropriate registry
changes so the client uses the local component, rather than looks for the remote component.
 Updating a Data Viewer 145

Figure 5.9
Using a local compo-
nent makes updating
the DCOM client easy.

However, this practice is for advanced programmers only—those with nerves of steel who
are willing to take the chance the application won’t work at all. Most developers will take Ch
the additional step required to update the reference and recompile the application.
5
Updating a Data Viewer
Many people interact with database applications on the Internet, but only to gather infor-
mation. A potential customer might want to look at the stock you have to sell or a partner
might want to learn the status of a project. Some people will just want to browse any free
information your company can provide. No matter what the reason, there are lots more data
viewer applications than full-fledged database management applications out there today.
While you may eventually upgrade the database management application used by employees,
it’s almost certain that your first step will be to create a data viewer.
Figure 5.10 shows the design diagram for a SQL Server database that I created. It’s a sim-
ple multitable example and we’ll explore it in more detail later. All you need to know now
is that this database contains a Client table that some of the sales staff want to look at
while on the road. I have a DCOM application that views the data in this table and we’ll
convert it to a SOAP application in this section. As with every other example in this chap-
ter, you’ll find the source code on the Que Web site for this book. You can find it at
www.quepublishing.com.

Before we actually perform the conversion, however, you might want to know a little more
about the three-tier programming concept. Many developers still use the client/server pro-
gramming model, but that won’t work well with SOAP. You need the additional features that
a three-tier application can provide. The remaining sections help you understand the
process for moving a data viewing application from the LAN to the distributed environment
of the Internet.
146 Chapter 5 Migrating an Application from DCOM to SOAP

Figure 5.10
A simple multitable
database provides the
basis for the data
viewer application.

Understanding the Importance of Three-Tier Programming

You could actually replace the three-tier in the title for this section with “multi-tier” or
“n-tier,” but most database documentation refers to it as the three-tier model. The three
tiers correspond to the client, business logic, and back-end processing portions of such an
application. The three-tier method of creating database applications is especially important
for SOAP developers because that middle tier provides the buffering needed for secure data
transmissions. You don’t want to expose your SQL Server setup to the world at large.
There’s a lot of misunderstanding about the role of three-tier development today and its
affect on the developer. In the past, a client would send data requests to a server. The server
would answer back. This is the client/server model. It’s not limited to two machines—there
can be as many servers as required to get the job done. The limitation is in the number
of levels of communication. When using the client/server model, the client requests data
directly from the server, which infers that the client knows the identity of the server. In
addition, the server can’t request data from another server on the client’s behalf—the client
and the server always have a direct relationship.
Unfortunately, client/server won’t work very well in the distributed computing environment.
As server farms become more prevalent on large networks, it’s not always possible for the
application to know which server will fulfill a request. Load balancing and other concerns
make it possible that one server will service a client during one request and another server
when making a second request. This is especially true with SOAP where each client request
is another session. In fact, it’s possible that a server will go offline and hand the request to
 Updating a Data Viewer 147

another server mid-way through the transaction. Given the realities of SOAP and the
distributed programming environment, direct contact is no longer feasible.
Another problem with client/server is that the client has to directly request every piece of
data it needs, which means the client has to have a lot of intelligence. Adding intelligence to
the client, rather than the server, means writing a lot of redundant code. In a distributed
environment, the client should only need to worry about making an initial request for data.
It’s up to the server to figure out how to get that data for the client. If the server has to call
on several other servers to fulfill the request, the client shouldn’t be aware of the back-
ground transactions taking place. Building an application using this technique means that
the server can handle a wider range of clients, all of which use a similar request format.
Consider, for a moment, that the developers of SOAP wanted to provide an easy to use pro-
tocol for distributed component use and discovery by a wide range of platforms. In this new
environment, anything from a PDA to a Linux server could access your system in search of
data. You can’t assume anything about the client, which means the intelligence for an appli-
cation must reside on the server. While the client is responsible for the presentation of the Ch
data (unlike in the days of dumb terminals), the server is responsible for gathering the data
in an easily recognizable (documented) format. 5
It’s important to understand that three-tier applications can work across many kinds of
boundaries because the request format is always the same. A thin client using a browser as a
container will request data using the same request format as a thick client operating on the
desktop. The server no longer needs to know what kind of client is making the request, only
that the client needs data and resources that the server can provide. In short, three-tier com-
puting and the distributed environment both require self-contained servers and clients, each
of which only knows about the data and objects that are being exchanged.
As you can see, three-tier computing isn’t about specific roles—it’s about modularity and flex-
ibility across vast distances. Unlike client/server, which vendors designed for use in a LAN
environment, three-tier applications are designed to work in the WAN or MAN environ-
ment, where you can’t assume anything about the connection between the client and server.

Updating the Common Business Logic Component

The DCOM application uses a simple server-side component in this case to serve up data in
the form of records. Here’s a sample of the code from the component.
Public Sub GetClient(ClientData As ADODB.Recordset)
‘Open the recordset.
deClient.rsClient.Open

‘Make the data accessible the client.

Set ClientData = deClient.rsClient

‘Close the recordset

deClient.rsClient.Close
End Sub
148 Chapter 5 Migrating an Application from DCOM to SOAP

You should notice an immediate problem with this code. SOAP can’t handle the recordset data
type. Another problem is that this component assumes the client will process the data. Of
course, that’s another problem with SOAP, you can’t assume anything about the client. A PDA
may not have the memory required to handle an entire recordset, much less manipulate it.
The component logic is simple in this case—there isn’t much you can do to simplify it.
However, the client side of the equation is quite complex, so that’s where you would look for
areas of simplification in this case. What we really need to do is leave the DCOM applica-
tion alone and create a new component that performs some of the work that we expect the
DCOM client to do. In short, we need to move some of the viewing logic from the client to
a new server-side component. Given the format of the data shown in Figure 5.10, we can
probably pass all of the information to the client in the form of strings after processing.

Separating the Data Viewing Logic from the Main Database Component
The SOAP component is actually a new server-side DLL in this example. Depending on
how you set your database application up, however, the server-side DLL could contain a lot
more than just viewing logic. We’ll need methods that return the next and previous records.
The client will keep track of it’s current position in the database by returning the Customer
ID value along with the next method invocation. We’ll talk in the “Updating a Complete
Database Application” section about the need to maintain state information for database
applications—this is one such technique.
Now that we have some of the preliminaries out of the way, let’s talk about some code.
Listing 5.1 shows the code that I created for the SOAP server-side DLL for this example.
Remember that most of this code used to appear in the client application; we just moved it
to the server.

Listing 5.1 A Server-Side Component for SOAP Use

‘Create the required components
Dim oOrderModify As ViewData.ViewClient
Dim rsCatalog As ADODB.Recordset

‘Create the data exchange variables.

Dim CustomerID As String
Dim FirstName As String
Dim MiddleInitial As String
Dim LastName As String
Dim Title As String
Dim Company As String
Dim Address1 As String
Dim Address2 As String
Dim City As String
Dim State As String
Dim ZIP As String
Dim Telephone1 As String
Dim Telephone2 As String

Option Explicit
 Updating a Data Viewer 149

Listing 5.1 Continued

Private Sub PerformUpdate()

‘Make sure we can display the record.

If Not (rsCatalog.EOF Or rsCatalog.BOF) Then

‘Gather the information.

CustomerID = rsCatalog.Fields(0).Value
FirstName = rsCatalog.Fields(1).Value

If Not IsNull(rsCatalog.Fields(2).Value) Then

MiddleInitial = rsCatalog.Fields(2).Value
Else
MiddleInitial = “ “
End If

LastName = rsCatalog.Fields(3).Value

If IsNull(rsCatalog.Fields(4).Value) Then
Title = “N/A” Ch
Else
Title = rsCatalog.Fields(4).Value 5
End If

If IsNull(rsCatalog.Fields(5).Value) Then
Company = “N/A”
Else
Company = rsCatalog.Fields(5).Value
End If

Address1 = rsCatalog.Fields(6).Value

If IsNull(rsCatalog.Fields(7).Value) Then
Address2 = “N/A”
Else
Address2 = rsCatalog.Fields(7).Value
End If

City = rsCatalog.Fields(8).Value
State = rsCatalog.Fields(9).Value
ZIP = rsCatalog.Fields(10).Value

If Not IsNull(rsCatalog.Fields(12).Value) Then

Telephone1 = rsCatalog.Fields(12).Value
End If

If Not IsNull(rsCatalog.Fields(13).Value) Then

Telephone2 = rsCatalog.Fields(13).Value
End If

End If
End Sub

Sub GetNext(sCustomerID As String, _

sFirstName As String, _
150 Chapter 5 Migrating an Application from DCOM to SOAP

Listing 5.1 Continued

sMiddleInitial As String, _
sLastName As String, _
sTitle As String, _
sCompany As String, _
sAddress1 As String, _
sAddress2 As String, _
sCity As String, _
sState As String, _
sZIP As String, _
sTelephone1 As String, _
sTelephone2 As String)

‘Open the recordset.

LoadData

‘Search for the current record.

Dim Search As String
Search = “CustomerID=” + sCustomerID
rsCatalog.Find Search

‘Make sure there is a next record.

If Not rsCatalog.EOF Then

‘Go to the next record.

rsCatalog.MoveNext
PerformUpdate

‘Get the data.

sCustomerID = CustomerID
sFirstName = FirstName
sMiddleInitial = MiddleInitial
sLastName = LastName
sTitle = Title
sCompany = Company
sAddress1 = Address1
sAddress2 = Address2
sCity = City
sState = State
sZIP = ZIP
sTelephone1 = Telephone1
sTelephone2 = Telephone2
End If

‘Close the recordset.

UnloadData
End Sub

Private Sub LoadData()

‘Instantiate the DCOM component.
Set oOrderModify = New ViewData.ViewClient

‘Gain access to the recordset.

oOrderModify.GetClient rsCatalog
 Updating a Data Viewer 151

Listing 5.1 Continued

‘Open the recordset.
rsCatalog.Open
End Sub

Private Sub UnloadData()

‘Close the recordset.
rsCatalog.Close

‘Clean up the objects.

Set oOrderModify = Nothing
Set rsCatalog = Nothing
End Sub

This code is somewhat truncated from the original on the Que Web site for this book. You
can find the full version at www.quepublishing.com. In addition to the GetNext() method,
there’s also a GetFirst() and a GetPrevious() method. All three methods work in essentially
the same manner. The one difference is that the GetFirst() method doesn’t need to search Ch
for a record.
Note the technique used to create the PerformUpdate() method. Every field that isn’t part
5
of the primary key or required in some other way has an alternate value. If you don’t include
code like this, then SOAP will send an error message to the client that it received data that
wasn’t marked “text/xml.” In fact, SOAP will often resort to this message when working
with database applications, leaving you without any clue as to the source of the problem.
Robust error handling isn’t an option when working with SOAP because it presents mes-
sages that are even more ambiguous than the normal variety.
The overall process for this code is that the method creates an instance of the original com-
ponent. This, in turn, opens the database for use. The method requests the data for the
current record, formats it as strings, and passes the information to the client. The method
call ends by closing the database and releasing locally created objects.

Creating the Data Viewer Client

The use of SOAP and an intermediate component simplifies the client. All the client needs
to do is create a connection to the server-side SOAP component, get the data in the form of
strings, and display it on screen. Listing 5.2 shows what the client code looks like for this
example (the code is again cut for the sake of brevity—check the Que Web site at
www.quepublishing.com for a full copy of the source code).

Listing 5.2 Data-Viewer Client Code

‘Create the SOAP client.
Dim Client As SoapClient

‘Create the data storage variables.

Dim CustomerID As String
152 Chapter 5 Migrating an Application from DCOM to SOAP

Listing 5.2 Continued

Dim FirstName As String
Dim MiddleInitial As String
Dim LastName As String
Dim Title As String
Dim Company As String
Dim Address1 As String
Dim Address2 As String
Dim City As String
Dim State As String
Dim ZIP As String
Dim Telephone1 As String
Dim Telephone2 As String

Private Sub cmdNext_Click()

‘Set up an error handler.
On Error GoTo ErrorHandler

‘Get the customer data and display it.

Client.GetNext CustomerID, _
FirstName, _
MiddleInitial, _
LastName, _
Title, _
Company, _
Address1, _
Address2, _
City, _
State, _
ZIP, _
Telephone1, _
Telephone2

‘Display the result.

DisplayResult

Exit Sub

‘Display a message when an error occurs.

ErrorHandler:

MsgBox Client.faultstring, vbExclamation

End Sub

Private Sub cmdQuit_Click()

‘Exit the Application
End
End Sub

Private Sub Form_Load()

‘Create the connection.
Set Client = New SoapClient
Client.mssoapinit _
“http://winserver/soapexamples/DCOMUpdate2/DataFormat.WSDL”, _
“DataFormat”, _
“ClientViewSoapPort”
 Updating a Complete Database Application 153

Listing 5.2 Continued

End Sub

Private Sub DisplayResult()

‘Display the result
txtCustomerID.Text = CustomerID
txtFirstName.Text = FirstName
txtMiddleInitial.Text = MiddleInitial
txtLastName.Text = LastName
txtTitle = Title
txtCompany = Company
txtAddress1 = Address1
txtAddress2 = Address2
txtCity = City
txtState = State
txtZIP = ZIP
txtTelephone1 = Telephone1
txtTelephone2 = Telephone2
End Sub
Ch
As you can see, this code is just an extension of the client code we viewed earlier. The main 5
differences are that I’ve placed some of the code in centralized procedures to accommodate
the needs of all of the methods. The cmdFirst_Click() and cmdPrevious_Click() methods
work precisely the same as the cmdNext_Click() shown in the listing. The code required to
create a connection is executed automatically when the form loads, which can save time.

Updating a Complete Database Application

In the previous two sections, you’ve learned how to convert existing DCOM applications.
We’ve used both an ASP and an ISAPI Listener approach in these two sections. These two
examples provide you with a good understanding of the DCOM part of the picture—what you
need to do in order to get an existing application running as quickly as possible under SOAP.
Chapter 8, “Providing Remote Database Access,” is going to show you how to create a full-
fledged database management example using a third-party product. In some respects, the
Microsoft SOAP Toolkit is still very much a diamond in the rough. I decided that another
toolkit is better suited to working with SOAP after a lot of consideration. While this is true
today, you may find that Microsoft has upgraded their product by the time you read this—which
means you may not need a third-party product to create fully functional database applications.

The Microsoft SOAP Toolkit doesn’t provide any means to validate the schema input
from a client against the schema within the WSDL file. This means that a client could
pass an invalid schema and you wouldn’t know that it was invalid until the server
failed. Unfortunately, a failure at this point masks the true source of the problem
unless you include special code within the component that manually validates the
schema. Microsoft plans to add validation to a future version of the toolkit. In the
meantime, you’ll need to check for schema related problems using additional code.
154 Chapter 5 Migrating an Application from DCOM to SOAP

This section of the chapter is going to look at some principles you need to understand in
order to convert a DCOM database application into something that will work with SOAP.
It’s important to consider how this kind of application differs from the other two applications
we have converted. Of course, certain principles remain the same, no matter what type of
application you create. For example, you’ll still need to verify that SOAP can handle the
data types in your application. It’s also important to check the WSDL file for problematic
conversions and inefficiencies.
A full-fledged database application sends and receives data. This two-way communication is
problematic when using SOAP because SOAP is a stateless protocol. Every time you call the
server, you’re starting from scratch. The server won’t remember anything you set up earlier.
This means that you’ll have to make every query self-contained. The query will need to
open the connection, make any required requests, obtain a result, and close the connection.
Of course, this makes using SOAP a lot less efficient than working with DCOM.
This is the strict interpretation of SOAP and you should follow it if you’re working with
more than one platform. Several vendors have concluded that some classes of SOAP appli-
cations will need to maintain state in some way. Database applications fall into the stateful
category. It’s inefficient to treat each query as a new conversation with the server. However,
since SOAP doesn’t know how to maintain state, you have to be resourceful in creating a
SOAP database solution.
It’s possible to maintain state outside of SOAP. You’ll have to jump through some hoops to do
it, but you can do it. Here are some tips for creating a stateful communication with SOAP.
■ Store the state information separately from the SOAP session. Some developers suggest
using a text file, others the registry, and still others Active Directory. The point is that
you need some external means of storing and retrieving the data.
■ Ensure you track the user’s information. The only way you can restore state informa-
tion is if you can associate the data with a particular session. The best way to do this is
to restrict the user to a single session, and then track the user identification informa-
tion. Since you’ll need to secure any database transactions over the Internet anyway,
you’ll always have user identification information at your disposal.
■ Provide some type of “watchdog timer” support for lost connections. DCOM and other
protocols periodically check the connection with the user by pinging the remote con-
nection. You can simulate this support by maintaining a local timer. Every time the user
makes a transaction, the application resets the timer. If the timer runs out, then applica-
tion rolls back any pending transactions and the state information is lost.
■ Include some type of transaction support. SOAP may actually provide this support as an
added feature by the time you read this. Developers currently use SOAP Actor support
to maintain transactions. A central server keeps track of the various transactions.
■ Maintain state-related code in a separate component. In many cases, you can use this
code for more than one component. The cost in coding hours of creating state mainte-
nance components is high, so you’ll want to get as much use out of the code as possible.
Make sure you use a modular approach that will work in a variety of situations.
 Modified Application Concerns 155

Once you have code that will maintain the user’s state, as well as provide transactional
support, you still need to consider the unreliable nature of the Internet. Working with the
Internet means accepting the possibility that some transactions won’t complete. Not only do
you need to maintain a strict multi-tier approach to coding your application, but also you
need to buffer the data in some way. It’s important to maintain the integrity of the main
database. Some developers rely on a buffering database that the server can scan on a regular
basis for errors. If the database integrity is maintained, the server commits the buffered
records in batch mode—incorporating them into the main database for use by others. Of
course, you’ll need to provide some level of training for users to ensure they understand this
approach to maintaining data integrity. The records an employee creates on the road today
won’t show up in the main database until after the next update period.
Some developers are reading this section and shaking their heads. It’s true that many devel-
opers view SOAP as the worst possible protocol for performing any sort of database work,
especially record updates and additions. Pure SOAP probably isn’t up to the job. You really
do need something that is completely reliable for mission critical applications. In some
respects, this brings us back to the DCOM application you’re using today. As mentioned in Ch
the very beginning of this chapter, you always need to consider that the protocol you’re 5
using today is the best one for the job. SOAP does add many new capabilities to the pro-
grammer’s toolbox, but it isn’t the end all of programming technologies.

Modified Application Concerns

I wish that I could say that every application you upgrade is going to work perfectly right
out of the testing lab, but you’d know that I was lying. Even applications that you create
from scratch and thoroughly test using a large beta group are likely to have problems. If you
ever doubt this statement, just ask Microsoft about the problems that they’ve experienced
over the years. If anyone should be able to create the perfect application, they certainly have
the resources and testing base to do so.
Converted applications are even more likely to have problems than ones you create from
scratch are. Not only do you have new code to consider, but you also have problems with
the assumptions made for the old application code to consider. In addition, updating an
application implies that you’ll spend time looking at code someone else wrote. Depending
on the skill of the other programmer, you may find yourself fixing their bugs as well as your
own. So, it’s probably a good idea to plan on spending time looking at problems with your
converted application when you first release it.
The following sections are going to look at the three major areas of modified application
failure that you’ll need to consider in addition to what we’ve talked about already: reliability,
security, and performance. If you’ve followed the advice in the previous sections, you’ll
already have considered problems with the protocol, code modifications in existing modules,
and code problems in new modules. You’ll have also considered some usage problems, such
as users who are well acquainted with one method of working with the application and
refuse to learn anything else.
156 Chapter 5 Migrating an Application from DCOM to SOAP

Reliability
Application reliability is one of the major areas of concern for a modified application.
DCOM is a two-way communication where the client and server have constant contact. The
client receives verification that certain events have taken place when working with DCOM,
so you can validate the exchange of data. In addition, you can easily add transaction support
to a DCOM application so that the client always has absolute verification that a certain set
of actions has taken place.
SOAP offers no such guarantees right now because it provides only one-way communication
and there is no guarantee that the server will respond to client requests. In addition, there’s
currently no way to create a transaction using SOAP, which means you no longer have the
option of checking the effect of a SOAP message exchange on the content of the database.
Microsoft and other vendors are currently working on remedies for this situation, but in the
mean time, you’re on your own. You could add code that would simulate the effects of using
transactions by using four message transfer stages for each transaction, but consider the sig-
nificant hit that your application would take in performance. The Internet isn’t the fastest
method to transfer data to begin with; adding an additional set of request/response messages
will only make matters worse.
One of the problems that I’ve run into quite often is that an application’s developer makes
certain assumptions about the presence or lack of error information. An application may
assume that it can verify a transaction between the client and server. It raises an error when
it can’t perform the validation, even if the transaction took place. The error handler might
further rely on application features that aren’t available when using SOAP, making it likely
that it, too, will fail. In the end, an application can fail, not because of any actual error, but
because of errors in the developer’s assumptions. We discussed the solution for error handler
problems in the “Error Handling Code” section of the chapter.
Some reliability problems that you’ll face are the result of using the Internet as a data transfer
media. The Internet is inherently unreliable—trusting the Internet to provide reliable data
transport is akin to trusting the postal system to deliver a letter. Generally, the postal system
gets a letter to its intended destination, but you have no guarantee this will occur and there-
fore need to accept some loss of reliability as part of the price of using the Internet. In short,
you need the transaction processing provided by DCOM, which is unavailable today. Until
the Internet becomes more reliable, you’ll need to add extra code to your applications that
detects errors that could occur because of a dropped line.
User unfamiliarity with Internet applications can also prove problematic. User training
resolves part of the problem by showing users the proper techniques for data entry and
other tasks. However, SOAP applications can create certain classes of problems that you
can’t avoid. Laptop keyboards are notoriously small and even a skilled typist will make more
entry errors. The character recognition features of PDAs can also prove difficult and you
need to check for new classes of data entry error that you may not have had to consider
before. In addition, users can get frustrated if they end up sending too many partial entries
to the server that are rejected later. A PDA screen is small in the extreme, making it likely
 Modified Application Concerns 157

that a user will click on what they think is an application feature such as help, when the but-
ton really sends the data to the server.
Until vendors devise a reliable transaction processing methodology, you should consider limit-
ing distributed application data sent by SOAP. Don’t bet the company on a protocol that isn’t as
reliable as the data transfer protocols you used in the past. Many distributed applications will
work better in mixed mode simply because older protocols like DCOM are more reliable.

Security
Modified applications present several security problems that you might not have had to
worry about with the original application. The most pressing problem is securing the data
transfer itself. This isn’t a problem when using DCOM because you can secure the applica-
tion using standard DCOM settings. DCOM offers a wide range of security features.
Remember from previous discussions, however, that SOAP doesn’t offer any security at all.
You need to add security as a separate feature, which means you can’t integrate the security
fully with the application. Any time you have a security solution that works separately from
Ch
the application, you open the possibility for security holes. You can’t guarantee a safe data
transfer because SOAP itself makes no such guarantees. 5
Creating dual mode applications also means checking two separate paths for security prob-
lems. Crackers love overly complex systems precisely because they are difficult to secure.
The dual mode nature of your modified application does reduce risk by isolating the external
modules from those used locally. However, the complexity that this configuration adds also
increases the potential for a failure due to security concerns.
Working with distributed applications means opening your network to the outside world.
After all, even if your network is secure, the user on the road is still a point of access that a
potential cracker could exploit. Consider the number of recent break-ins where an employee
accessing a corporate network from home opened the door for a cracker invasion. In fact,
the security of remote machines has become a major issue. Many companies are looking for
ways to increase their network security by securing these home machines.
The modified application is also a slave to the security provided by the Web server. SOAP
accesses your network through an Internet connection that the Web server handles. If the
Web server is less than secure, then so is your application. This problem is severe enough
that the Federal Bureau of Investigation (FBI) recently starting issuing warnings to major
corporations in danger of compromise by crackers. The problem is simple, these companies
haven’t applied all of the required security patches to their severs. The bottom line is that
the security of your modified application is now in the hands of a network administrator
who may not apply all of the patches required to keep the Web server safe.
Your own code is a potential source of security problems. Assumptions made by a previous
developer can affect the operation of your application. The code you use to make connec-
tions might assume that there’s a secure DCOM connection in place. It might not have the
additional code required by SOAP to make the connection secure. In short, if you simply add
the code required to make the application work, you might be opening a huge hole in security.
158 Chapter 5 Migrating an Application from DCOM to SOAP

Performance
As part of the process of modifying an application, you need to ensure the application makes
use of all modern performance enhancing features. DCOM has been around for a long time
so there’s a good chance that the application you modify today will use the programming
techniques of yesterday. Those techniques didn’t always consider performance because the
performance feature may not have existed.
The choice of object pooling and threading options is important because SOAP applications
rely heavily on them. Anytime you can service user requests from cache, rather than making
another call to the hard drive, you have increased the performance of your system. Likewise,
calling on multiple servers almost always results in a performance hit. It’s important to keep
the number of calls to a minimum whenever possible.
Remember that we modified the organization of the application at the beginning of this
chapter. The purpose of this change is to allow internal and external call separation to
ensure data integrity and application reliability. Using a segregated model also improves
security. Whenever you make changes of this sort you unfortunately complicate the perfor-
mance picture. Even the DCOM version of the application might not work as quickly after
the change simply because you need to call more modules to complete any given task.
Measuring performance is difficult at best. A SOAP application doesn’t have the same connec-
tivity that a DCOM application does, so you need to consider a myriad of factors when testing
performance. You can break down the performance factors for a SOAP application as follows:
■ Client message creation
■ Transmission
■ Web server handling
■ XML parsing
■ Server processing

The three factors that affect performance the most are transmission, Web server handling,
and XML parsing. You need to factor out the transmission time by performing a comparison
with utilities such as Ping. The transmission time will vary significantly given line conditions,
Internet loading, number of hops, and server loading. Retaining this factor in your perfor-
mance figures gives a false indication of application performance by making the application
appear slower than it actually is. However, you should factor in an average expected trans-
mission time and make sure that any reports you create provide realistic best and worst-case
scenarios.
Web server handling delays often occur due to improper configuration. For example, most
servers offer a keep-alive setting. Not using this particular feature means the server creates a
new session every time the client makes a request—wasting valuable time. In addition, you
need to check for bandwidth throttling and any TCP delay settings on the socket. All of
these configuration issues will cause delays with your application that will make it perform
slowly when compared to DCOM.
 Troubleshooting 159

The XML parsing issue is problematic. Many developers complain that the current “off-
the-shelf” XML parsers provide abysmal performance characteristics. One developer has
gone so far as to write a custom XML parser with better performance characteristics
(http://www-106.ibm.com/developerworks/xml/library/x-elexml/index.html). The trade-
off of using a custom XML parser is that some SOAP toolkits will break or not provide the
anticipated results. For example, the Microsoft SOAP Toolkit relies on their XML parser
version 3.0 or above. Even the older versions of the XML parser won’t work with this toolkit.
The client and server processing delays are the factors most under your control. Creating
components that perform tasks in an optimized way is the best insurance you have for pro-
viding good application performance. Make sure you allow for external setting of any delays
within the component so that the administrator can tune your component to reflect current
conditions. Validate client and server performance locally using a LAN so external factors
such as transmission times don’t affect the tuning process.

Troubleshooting
This chapter has shown you how to upgrade existing applications to use SOAP. In many
cases, you won’t convert the application completely; you’ll add SOAP as a new feature to the
existing application. Even if you do convert the application completely, you can still make
use of the code you already have in place and upgrade the techniques you use to perform
tasks such as handling errors. The following sections contain some of the questions that
developers seem to have about upgrading their current applications (based on newsgroup
and list server input).
Which DCOM applications are the best candidates for conversion?
This is one of the harder questions developers have to answer because it depends on many
factors. Sometimes the determining factor is more related to the money a company has to
spend on the upgrade, than to the technical feasibility of the project. When looking at an
upgrade project, you should always consider the following issues:
■ Do you have all of the required source code?
■ Did the originator document the source code well?
■ Do you have all of the accompanying documentation so that you can create a block
diagram of the application?
■ Will it require more time to upgrade the project than create a new one from scratch?
■ Is SOAP even the right solution for this distributed application project?

Will a SOAP application perform as quickly as its DCOM counterpart?

Of course, the real question in many situations is whether the DCOM application will work
at all in the distributed environment. One of the biggest reasons for the creation of SOAP is
to allow applications to work across firewalls and other Internet security measures. In some
cases, you may not have any choice other than to use a technology such as SOAP to upgrade
160 Chapter 5 Migrating an Application from DCOM to SOAP

your DCOM application to work in a distributed environment. However, you can also
summarize the performance factors for a SOAP application as follows:
■ Message size
■ Availability of transmission bandwidth
■ Security requirements
■ User requirements
■ Performance of any new client platforms

What are the interoperability considerations for a SOAP application versus those of DCOM?
DCOM applications work on the Windows platform and that’s it. When you reduce the
number of platforms that a technology will work on, you also reduce the number of poten-
tial interoperability problems. SOAP will work on multiple platforms. As a result, you’ll face
more interoperability problems when using this technology.
The number one interoperability problem is the data types used by the component versus
those that SOAP supports. Appendix A will fill you in on the details. Anyone who is familiar
with XML knows that it only supports certain data types and doesn’t support some of them
the same way that the native platform does. SOAP supports a subset of the XML data types,
so you’ll likely spend time figuring out ways to convert existing component inputs and out-
puts to something that SOAP can understand.
Another area of concern is vendor support. We talked about this issue a little in Chapters 1
and 2, but it bears repeating here. You need to consider the capabilities of the SOAP toolkits
you need to use as part of the interoperability problem. For example, some toolkits support
WSDL files, while others don’t.
 CHAPTER
6
Creating Remote Access Utilities
In this chapter
Introduction 162
An Overview of Remote Access Utilities 162
Writing a Server Status Viewer 172
Creating a Simple Employee Check-In Application 182
Project 193
162 Chapter 6 Creating Remote Access Utilities

Introduction
Some of you might have assumed that everything we’ve talked about so far deals with some
type of remote access. Actually, the examples in the previous chapter will work just fine in
either a local or remote access scenario. You can use the previous examples equally well on a
LAN and the Internet. Of course, many SOAP applications will emphasize remote access
because that’s one of the main reasons to use SOAP.
This chapter will emphasize remote access utilities. These are smaller applications that you
would normally use an application to perform locally. For example, we’ll look at a server
status application in this chapter. Normally, you’d use an administrator utility, such as
Microsoft Management Console, for gaining access to server status information locally.
Getting server status information can prove difficult from a remote location; using a SOAP
application can reduce that problem. Using a SOAP application will also allow you to share
some types of information with those outside your company without revealing all of your
company’s secrets. A server status utility will also provide a service for customers who want
to know whether your system is ready to receive orders or perform other work.
The first section of the chapter, “An Overview of Remote Access Utilities,” discusses several
remote access utility development issues. We’ll talk about how you can use remote access
utilities for your business, some of which I’ll demonstrate in the application programming
examples. This section will also tell you about working with Web Services. Many businesses
consider Web Services the next “killer application.” Finally, we’ll talk about development
issues, such as application flexibility, using existing components, security, and other prob-
lems that you normally don’t need to worry about.
The second section of the chapter, “Writing a Server Status Viewer” is an administrator tool
example. We’ll discuss how you can use SOAP to create remote monitoring software. Of
course, one of the more common monitoring requests is status information.
SOAP applications also need to support the user on the road. The third section of the chap-
ter, “Creating a Simple Employee Check-In Application” shows one user-support technique.
This is a relatively simple, but common task that businesses perform. We’ll look at other
applications that support the user on the road as the book progresses.
Finally, the “Project” section of the chapter will lead you through the process of creating a
developer utility of your own based on all of the information we’ve accumulated so far. The
purpose of this section is to provide you with hands-on time with this new technology. This
section is less example-oriented—more hands on training time.

An Overview of Remote Access Utilities

What is a remote access utility? That’s a good question to ask because many people have a
preconceived notion of what this type of utility is without any basis in fact for their opinion.
Many developers view remote access utilities as small applications that perform a single task.
It just happens that the utility also performs this task across the network wire, making a
remote access utility rather than a local utility. Other people view remote access utilities as
 An Overview of Remote Access Utilities 163

some type of new application. Some developers go so far as to equate remote access utilities
with Web Services. In short, we need a working definition of remote access utilities before
we can even explore the genre in light of SOAP application development.
The following sections will help clear away some of the confusion. We won’t define a
remote access utility as a single application type, but as applications with specific characteris-
tics. Many of you will recognize your personal definition in this section, but that definition
will also appear with other utility types you might not have considered. The concept of
remote access utilities is broader than you ever thought. Certainly, I was surprised at the
diversity of applications that people came up with when questioned about the topic.

Uses for Remote Access Utility Applications

It’s important to understand that SOAP, like any protocol, has limitations that you need to
observe. For example, one question that I saw online was how to pass a java.awt.Image from
server to client using SOAP. The response to this question was that if you want to pass a
java.awt.Image, you should use Java Remote Method Invocation (RMI). Using Java RMI
allows the developer to use native coding techniques. In short, SOAP isn’t the end-all
answer to every programming need, even when creating a utility.
Of course, the problem of passing an image from a server to a client has a SOAP solution.
One of the suggestions is to pass the image as a Graphic Interchange Format (GIF) byte
array. If you look at Appendix A, “SOAP Data Types and Data Type Conversions,” you’ll
find that this is one of the native formats for SOAP because it’s also a native format for
XML. Some developers suggest using a data stream for the job and allowing the SOAP
server to format the message as a Multipurpose Internet Mail Extensions (MIME) attach-
ment. The point is that you need to use SOAP in the way that the designers intended.
This leads to the discussion at hand: how you can use SOAP for remote access utilities.
Downloading graphics images is one example of a remote access utility. You don’t need a
large program to download the graphics. The only time you need a large application is if
you want to allow the user to manipulate the graphics in some way. This differentiation is
one of the things you should consider for a utility program. A utility normally performs sev-
eral simple functions, some of which may be a subset of what you would get when working
with a larger application. The following lists some of the best uses for utility programs when
working with SOAP: Ch

■ Monitoring: One of the most important classes of utility application is the type that 6
monitors some type of activity. Microsoft has stuffed Windows with utilities that per-
form this service. Many network administrators also know Unix for the utilities that it
supports. Some of these utilities operate at the command line, while others support a
GUI. Many support some type of remote operation to allow a local use to check the
status of another machine on the network. SOAP can extend the capabilities of these
applications from the LAN to the Internet.
■ Simple Data Manipulation: We’ve already talked about one form of this remote access
utility. Manipulating graphics in some way is a task that most developers will perform at
some time in their career. The simple act of moving a graphic from the local network to
164 Chapter 6 Creating Remote Access Utilities

the Internet can prove problematic without the proper protocol support. SOAP pro-
vides a method for smoothing the movement of data, but it can also allow you to access
and manipulate the data from a remote location.
■ Services: Developers usually call utilities that provide access to some type of resource a
service. These utilities do everything from monitoring the number of users accessing
your server to updating the time on all of the network workstations. The Internet has
extended the concept of service from the network to the world. SOAP provides the
means for accessing these Web services. However, you don’t have to reserve SOAP for
those services that everyone will access—SOAP works fine on the LAN as well.
■ Configuration: Utilities often allow users to configure an application, an operating sys-
tem, a component, or another resource. In some cases, utilities allow centralized config-
uration of many resources. For example, developers commonly use a single utility to
configure all of the hardware on a machine. In other cases, a single utility will allow
configuration of many different objects. For example, operating systems often handle
security configuration using a single utility. SOAP simply allows you to extend your
ability to configure resources from the local workstation or network to the Internet as a
whole.
■ Maintenance: Keeping a system running properly requires maintenance. However, per-
forming maintenance on a large system is often a painful task because utilities might
require local access. In recent years, the use of agents has allowed remote utility use,
but primarily over the network alone. SOAP will extend the maintenance utility by
allowing the network administrator to gain access to low-level information across the
Internet.
■ Task Scheduling: Automation of all types is a requirement in today’s harried world of
overworked network administrators. Most operating systems include a simple task
scheduler that works for most situations. However, developers still have to create cus-
tom schedulers for complex applications. The scheduler often takes the form of a utility
that resides outside the main application.
■ Other: Only your imagination limits the ways in which utilities appear within your
company. You’ll find that many developers create and use utility applications without
much thought. They represent the small, single task applications that someone created
for a simple need. SOAP merely extends the ways in which you can use the simple
utility within your organization and those outside it.

Remote Access Applications—Not Necessarily Small

Most of the remote access applications you run into will provide simple, single-task functionality. In many
cases, remote access applications are also small and easily described. However, these two features describe
only most of the applications out there. A few remote access applications are quite large and conceivably com-
plex in implementation, albeit simple in concept.
Consider the current effort by Ariba, IBM, and Microsoft to develop the first truly functional “yellow page” ser-
vice for the Internet. The Universal Description, Discovery, and Integration (UDDI) registry will provide a tele-
phone book-like service for businesses worldwide. Most people would agree that a telephone book is relatively
simple in concept, but complex in implementation.
 An Overview of Remote Access Utilities 165

It seems that this UDDI effort is a little more robust than the average telephone book. Businesses will actually
participate in three types of telephone book entry: white, yellow, and green.
A white page book will contain general business information, such as name, address, telephone number, e-mail
address, and Web site URL. The white page book will also contain information about the types of services that
a business provides and which protocols it supports. As with most white pages, the supporting vendors will
organize this one in alphabetical order by business name, but given the realities of computer flexibility, you’ll
probably be able to sort it in any order needed.
The yellow page book will tag businesses with government operation codes. These codes standardize the business
classification, such as dry cleaners or consulting services. The vendors intend to organize the yellow page book by
operation code, geographic area, international naming protocols, and technology-based naming protocols.
The green pages are where this book comes into play. They describe what types of documents a company can
receive, entry points for transactions, and the types of technology the business supports. This information will
help you find companies that support the same documents that your company does. It means that you’ll be
able to write a single client that can interact with multiple companies, even if you don’t know what protocols
that company will support now. In short, this type of remote access application will save time in the future.

Understanding the Web Services Difference

I mentioned Web Services several times in the previous section. One of the biggest reasons
to use SOAP with a utility is to provide resource access to those outside of your company.
The resource could be something as simple as the local time or as complex as your com-
pany’s catalog. The utility could provide one-way access in the form of an informational dis-
play, or two-way access in the form of a data manipulation program. The bottom line is that
the utility provides a service that people outside your company might need.

Don’t get the idea that Web Services necessarily equate to component technology. A
Web Service can be any resource that a vendor wants to make available to the public
at large using a standardized access method. In addition, Web Service access doesn’t
have to rely on SOAP; you can access Web Services using other techniques. This sec-
tion of the chapter is looking at one possibility for Web Service support. A discussion of
Web Services as a whole is outside the scope of this book. We’re looking at a specific
view of Web Services.

Ch
One of the most obvious ways in which a Web Service differs from the utilities you’ve cre-
ated in the past is that people outside your company will need to access it. This means that 6
you’ll want to polish the appearance of the user interface a bit more than usual and might
want to provide some form of help with the service (if possible). Security is also an essential
part of a Web Service. You want to ensure that people using Web Services won’t gain access
to parts of your company that you want to remain hidden. Reliability is also a concern. Your
customers won’t be impressed if the service your company provides constantly breaks.
Web Services also differ from the typical utility in that you’ll need to advertise them in some
way. If you only want selected customers to know about the utility, then a simple phone call
or e-mail announcement will suffice. However, if you’re like most Web Service developers,
166 Chapter 6 Creating Remote Access Utilities

you won’t even know the people using your service—at least not immediately. This is where
other protocols, such as UDDI, come into play. These other protocols help make your
SOAP applications visible to others.
Some vendors, such as Microsoft, are tying Web Services to their applications. The applica-
tion will theoretically work fine without the Web Service, but adding the Web Service
increases application functionality in some way. By charging a small amount each time a user
accesses the service, the vendor can create a continuous revenue flow. In this case, the Web
Service is a value-added packaging methodology.

Web Services is becoming an increasingly big issue as companies seek ways to connect
through the Internet. You’ll find many viewpoints about this technology. Get the IBM
view at http://www.alphaworks.ibm.com/tech/webservicestoolkit. The
Microsoft perspective appears at http://msdn.microsoft.com/webservices/.
You can find a wealth of discussions at http://search.userland.com/
default?s=1&m=50&q=Web+Services&site=All.

At least a few vendors also see the value of Web Service for collaboration. For example, a
travel agent could add your hotel reservation component to their application. Every time
they need a hotel reservation that your company can fulfill, your component makes the
required entries in the database. Using Web Services this way means that companies can
create extremely efficient and complex applications using component parts. The best part is
that the companies that use your Web Service don’t need to know anything about your data-
base or the method used to access it. All they need to know is the standard document format
that your company uses for exchanging data.

When Is SOAP Overkill?

Some developers will ask whether they need to use SOAP for remote access utilities at all.
The answer can become complex. A short answer is that it depends on what you expect the
application to do when complete. SOAP truly is overkill for some types of applications,
especially if your sole purpose in implementing the SOAP application is getting around a
pesky firewall.
For some experts, the keyword for SOAP applications is interoperability. If your application
will never see use outside your company, then a solution that relies on ASP might be all you
need. It doesn’t pay to use SOAP if a simpler technology will work. However, it’s hard for
most developers to determine at the outset of a project whether outside entities will use the
data from that project. Determining interoperability concerns at the outset of a project is
important, but we all know those concerns will change as the project develops.
Other experts see the question as one of complexity. SOAP allows you to overcome specific
types of obstacles when working with Web applications. Although an ASP-based application
might work for viewing static data, these experts argue that modern applications require
greater flexibility. SOAP can provide much needed flexibility in Web-based applications.
However, if all the user will ever do is view data, SOAP could be overkill. You’ll normally
 An Overview of Remote Access Utilities 167

reserve SOAP for situations in which some type of data exchange (one-way or two-way)
takes place.
Another concern for remote access utilities is longevity. SOAP does require an investment in
time and effort. Although this investment will become less as SOAP toolkits mature, you
still need to consider the investment. You’ll waste that investment if the company uses the
application for a short time and then discontinues in favor of a new solution. Utility pro-
grams are risky in this respect. Many people begin developing a utility as throw-away code
for a task of the moment. In some cases, the utility outlives the originally intended use. In
many cases, though, it dies a quick death when the developer finishes using it.

Making Utility Programs Flexible

One of the ways to hedge your bets with utility programs is to make them both modular and
extensible. Modular utilities tend to perform one task well per component. Modifying such a
component to perform another, similar task is usually simple. In fact, if you design the com-
ponent well, you might be able to use it for multiple tasks without modification.
Extensible components contain code that allows augmentation of a particular function. For
example, a configuration setting might adapt the component to other purposes. A basic
math module might perform different tasks depending on the equation that you feed it as
part of the input. (This is obviously a simple example, but one that clearly illustrates the
point I’m trying to make.)
You can create configuration settings for components in a number of ways. Here are the
techniques I use from least to most difficult.
1. The easiest method is to pass a configuration string as one of the arguments, but this
isn’t always practical and can cause security problems.
2. The second method is using registry entries that the component can check during the
initialization phase. This method is a little more secure, but tends to get messy when
you want to configure different instantiations of the same object in different ways.
You’d need a registry entry for each potential configuration and some method for dif-
ferentiating between them.
3. You could create the component as a COM+ component and place the configuration
settings in the COM+ application or within the COM+ Catalog. This method is Ch
more secure than the registry method because the settings remain on the server. It is
also less messy when you need to configure the same component in different ways.
6
Using COM+, however, adds complexity to the application that you might not want
to support.
4. Placing the settings in a stream and saving them to disk is another method. The com-
ponent can load the property bag and use the settings it contains to configure itself dur-
ing the initialization phase. This method is less secure than COM+, but more secure
than the registry method because the settings are still located on the server. The meth-
ods for working with a property are well documented and precise. However, some
developers find that this technique has all of the problems of the registry method with-
out the ease of access that the registry provides.
168 Chapter 6 Creating Remote Access Utilities

Shortcuts for Using Existing Components with Utilities

We all have “junk” components sitting around. They were useful at one time, but they’re
not usable today for whatever reason. Most of us avoid creating new junk components
because of the time wasted. Creating and debugging a component takes time and effort that
we could use in other ways. Developers are in a constant time crunch, and the situation only
promises to get worse. The big question, then, is how to use existing components within
SOAP applications.
The surprising news is that some components will run just fine as they are right now.
We’ve already discussed this fact as part of the Distributed Component Object Model
(DCOM) discussion in Chapter 5, “Migrating an Application from DCOM to SOAP.”
Converting old applications isn’t only possible, but it’s also the recommended method by
Microsoft and other vendors. Large vendors want to encourage code reuse so that you’ll
try their new technology. It’s in their best interest to make sure you can use as much code
as you can.
I’ve actually managed to resurrect some aging components from the scrap heap by using
SOAP. This might seem somewhat amazing until you consider the fact that we’re taking a
half step back in the component technology development. All SOAP toolkits support DLL
(in-process) components—the same components you’ve been using on the desktop for ages.
Remember that out-of-process (EXE) components didn’t really become popular until
DCOM arrived on the scene, and developers had to make the switch.
Some, but not all, SOAP toolkits also support the EXE form of component. Unfortunately,
the Microsoft SOAP Toolkit isn’t one of those that do support the EXE format. If you’re
using Visual Basic, switching to a DLL format might be as easy as loading the project,
changing the Project Type setting to ActiveX DLL (see Figure 6.1), and recompiling the
application. However, what happens if you don’t have the source code for the component or
the component isn’t easy to change to a DLL format? Many developers create a DLL com-
ponent that loads and accesses the EXE version. This method has the disadvantage of
increasing latency and using more server resources. This might not be a problem if only a
few users require access to the component and the component is small.
The “second tier” component approach also works for cases in which the old component
uses data types that SOAP doesn’t recognize and you don’t want to go to a lot of effort to
rewrite the code. You can use the second tier component to translate the data types from
SOAP into something the older component will understand. Of course, now you’re adding
still more latency and resource usage to the picture.
Another approach to handling incompatible data required by an older component is to
write a user-defined type (UDT) for SOAP. In this approach, you’ll create the Web
Services Description Language (WSDL) file as usual. It will contain areas with question
marks where the utility didn’t understand the data type that the component used. You’ll
need to edit the file manually to define the unrecognized data type and provide code to
handle the data transfer.
 An Overview of Remote Access Utilities 169

Figure 6.1
Sometimes, switching
the project type is all
you need to do.

Utility Program Security Issues

Utilities have all of the same problems that most applications do when it comes to security.
Any data you transfer is open to scrutiny until you secure it; users are just as apt to dig for
hidden secrets as they are when working with a database application. In fact, you might find
that users are even nosier than usual because they’ll assume that your utility program isn’t as
secure as the big application.
Utility programs do have an advantage in that the data they handle is normally less sensitive
than your typical database application. In addition, utilities normally transfer small amounts of
data, making it harder to intercept something that you wouldn’t want someone else to know.
Some security issues are peculiar to utility programs. I’ve already mentioned one potential
security issue for utility programs: setup strings. Crackers will use even the smallest cracks in
your security to cause problems. In fact, they tend to like small, unnoticeable problem areas.
Ch
That’s why you’ll want to be sure that if your utility relies on configuration settings, you act
to secure them. In some cases, the settings might be so innocuous that you don’t have to do 6
anything at all, but this is likely to be the exception, rather than the rule.
The fact that utility programs often provide low-level access to your server or the applica-
tions it supports is also a problem. I know that I wouldn’t want someone outside of my orga-
nization to use some of the utilities that I create. Sure, they provide useful functionality, but
that functionality is such that the security of my system would be at risk if someone else
accessed them. Not all utilities have this problem, but you’ll still want to consider the issue
when you create the SOAP version.
170 Chapter 6 Creating Remote Access Utilities

Of course, the biggest threat to security when working with utility programs isn’t the
cracker lurking outside your firewall with evil intent—it’s the employee working freely
inside the firewall whom the boss fired just a moment ago. Directing your attention outside
and forgetting the inside of your organization is asking for trouble. Utility applications are
small, fast, and specifically designed to perform one task well. Someone bent on destruction
after a bad day in the office won’t have to go through much effort to wreak havoc on your
system. The very feature that makes utility programs a must-have also makes them incredi-
bly dangerous for the unwary developer.

Non-Issues for Utility Programs

The simple nature of utility programs means there are issues you don’t need to consider. For
example, it’s unlikely that you’ll manipulate a database with your utility program—that’s
normally the domain of larger database applications. As a result, you’ll find that you don’t
have to have multiple tiers of components to process requests reliably. Utility programs are
simple.
The following sections will help you understand some of the non-issues for utility programs
a bit better. These are issues that you’d normally think about when creating an application,
but that you can disregard when working with utility programs. The first non-issue is a spe-
cial SOAP header processing consideration. Utility programs normally provide direct access
to the data they manipulate, so you don’t have to worry about actor support. We’ll look at
other issues as well.

SOAP Headers Processing

One of the issues that developers worry about is how SOAP will treat SOAP messages with
multiple header entries. Each entry defines a special processing need. You might want one
server to check the reliability of the message and another to check for any log entries associ-
ated with the message. For example, look at this multiple entry SOAP header. (This header
is by no means typical—I’m only using it for discussion purposes.)
<SOAP-ENV:Header>
<a:authenticate xmlns:a=”uri1”
SOAP-ENV:actor=”http://www.servera.com/authenticate”
SOAP-ENV:mustUnderstand=”1”>
....
</r:authenticate>

<t:transaction xmlns:t=”uri2”
SOAP-ENV:actor=”http://www.serverb.com/transaction”
SOAP-ENV:mustUnderstand=”1”>
....
</t:transaction>

<l:log xmlns:l=”uri3”
SOAP-ENV:actor=”http://www.serverc.com/log”
SOAP-ENV:mustUnderstand=”1”>
....
</l:log>

</SOAP-ENV:Header>
 An Overview of Remote Access Utilities 171

The details of this header aren’t important. Consider for a moment, however, that different
servers process the authenticate, transaction, and log entries. Server A will process the
authenticate header and pass the resulting message to Server B. Note that we’re talking
about the resulting, not the original message. Server B will process the transaction header
and pass the resulting message to Server C. Server C will process the log header and pass
the result to the client. In short, SOAP relies on a processing change, not a centralized pro-
cessing method. This differs from technologies such as COM+, where a central authority
monitors process.
What happens if Server B experiences an error? Instead of passing the message to Server C,
Server B will report the result back to the client. Header processing always stops after the
first error. This means the client gets immediate feedback and the system doesn’t waste time
processing a request that won’t succeed anyway. In some ways, this makes SOAP more effi-
cient than some technologies such as COM+, where all servers perform the requested task
and then vote on the outcome of a transaction.
This methodology has implications for the developer as well. It means that you’ll end up fix-
ing one error at a time. Even with proper development tools, the fact that SOAP stops pro-
cessing on the first error means that you’ll fix one error at a time when debugging an
application. Any developer who has spent much time working with distributed applications
knows that the one-at-a-time approach is the most time-consuming method to debug an
application. Given the way that SOAP handles headers, you’ll want to set additional time
aside for debugging when working in this environment.

Interoperability
We’ve spent a lot of time in this book discussing the issue of interoperability. From the
beginning of Chapter 1, “An Overview of SOAP,” to the last example in Chapter 5,
“Migrating an Application from DCOM to SOAP,” interoperability has been an issue
because SOAP is a new protocol. The fact that servers from different vendors can communi-
cate at all is a miracle considering that this never occurred when working with DCOM or
CORBA. Microsoft never ported DCOM to other servers, and you’d be hard-pressed to
find CORBA on every platform either. The imperfect implementation of SOAP we have
today is definitely a step in the right direction, but it will take time for vendors to make
things work correctly. Ch
You don’t have to worry about interoperability issues when working with utility applications 6
for the most part, unless your organization sports dozens of incompatible platforms. People
within the organization use utility programs that actually span the network, not those out-
side of it. Keeping the application for local use means you don’t have to worry about outside
clients attempting to access your component.
Utilities that are used by outside parties tend to use simple data types. For example, some-
one who wants to configure his or her user settings will normally send strings to your com-
ponent, not complex data. Every SOAP toolkit on the market today supports the basic data
types listed in Appendix A, so it’s unlikely you’ll run into interoperability problems in this
arena either.
172 Chapter 6 Creating Remote Access Utilities

Attachments
Remember that a SOAP message can appear as part of a Multipurpose Internet Mail
Extensions (MIME) message. The other parts of the message can contain attachments that
don’t transfer well using XML encoding, such as graphics. Utility programs rarely, if ever,
need to transfer complex data of this type, so you seldom need to worry about attachments.
Using the standard high-level messaging technique found in Chapter 4, “Using SOAP to
Create a Simple Application,” makes creating messages much easier and faster.
Obviously, there are exceptions to the rule, but you can normally reduce the amount of cod-
ing you need to perform. For example, consider the simple data manipulation entry in the
“Uses for Remote Access Utility Applications” section of the chapter. This is, in fact, the
only class of utility for which you’ll need to consider attachments. You might use a utility to
move a graphics image from the client to the server. Creating generalized code to perform
the task will allow you to write the application once and never worry about it again.

Experimentation
Some developers I know are stressed about every line of code they write; they don’t have fun
with their trade (in part, because they’re always under severe deadlines). Utility programs
are my favorite kind of coding because they allow me to experiment. You can’t get away with
writing a million-line database application and then throwing it away. Everyone will think
you’re crazy.
On the other hand, I commonly write small utility programs to test a new idea. I’ve discov-
ered a lot about programming by sitting down and seeing what’s possible with a utility. The
programs are small and no one cares if I throw them out later—not even me.
This concept of toss-away code is foreign to many developers today because everyone is
thinking of ways to save every line of code. Writing lots of code is expensive; it’s an invest-
ment worth protecting. However, utilities are cheap when it comes to coding time. You gain
knowledge, and if you design the utility properly, you perhaps have a new piece of code to
add to your library.
The final non-issue for utility programs is the time invested coding them. The utility pro-
gram is small, easy to understand, and performs just a few tasks well. The education you
receive more than pays for the time you invest in creating one. The functionality you get is
just icing on the cake.

Writing a Server Status Viewer

This example will extend the example in Chapter 5 by showing another way of gathering
server status information. We’ll add the ability to poll the status of all of the services
installed on the server. Of course, you can make this application as robust as you want; this
example provides a skeleton you can build on later. The important consideration in a multi-
platform environment is finding statistics that every server supports, or at least customizing
the client to automatically compensate for platform differences.
 Writing a Server Status Viewer 173

The server-side component in this example works with the Windows API to poll the state of
each installed service. You’d need to provide a different server-side component for every
platform you want to support, but you can use the same component for all servers with the
same operating system, such as all Windows machines. The client, however, could remain
the same because all we’re passing is strings from the server to the client. As long as you
maintain a consistent interface, you won’t experience problems with using the same client
across multiple platforms. The servers, of course, will present interesting challenges because
each vendor has a different way of measuring the current server state.

Creating the Server-Side Component

The simple part of this component is that it requires no input and provides a string as output.
The server-side component for this example performs low-level access of the operating system
using API calls to complete the task. You’ll find that this is common for certain categories of
utility components. A utility provides a “black box” for an outside application that makes it easy
to perform system-level tasks. In this case, we’re polling the system service state on a server.

Polling a server for service information is relatively easy when you’re working with Visual
C++, but presents some interesting challenges when working with Visual Basic. The
ServicesDeclarations.BAS file found in the \Chapter 06\Service State Component directory
of the source code available from the Que Web site will help make working with services in
Visual Basic much easier. You can find the source code at http://www.quepublishing.
com. The example component shows how to use the various service functions to obtain
state information. You can also remotely add new services, delete services that you no
longer need, and modify the state of existing services using the same techniques.

Another criterion for this component is that it performs this service in a server-neutral fash-
ion. In other words, you should be able to place the component on any server and it should
work without any configuration. In some cases, this would be a difficult requirement because
many system calls require server-specific information. Fortunately, the API calls we’re using
in this example don’t require precise server knowledge, so creating a generic component is
relatively easy.
Now that you have some idea of what we’re going to do and why we’re going to do it, let’s Ch
look at the code for the server-side component. Listing 6.1 contains the source code for just
the component class. It’s important to remember that the component requires a lot more code
6
to function in this case. The additional, generic, code appears in the ServicesDeclarations.
BAS module as part of the source code available from the Que Web site.

The source code makes every effort to use API standard names for function calls and
constants. You can look up these functions in any Windows API help file for a full dis-
cussion of all arguments and constant values. The DLLs do use alternative names for
the functions to support both standard and wide (Unicode) character sets. You’ll find
these alternate names in the ServicesDeclarations.BAS file.
174 Chapter 6 Creating Remote Access Utilities

Listing 6.1 GetStatus() Method Source Code

Public Function GetStatus() As String

Dim SCM_Handle As Long ‘Result value or service control

manager handle.
Dim BufferSize As Long ‘Returned service lising buffer size.
Dim Result As Long ‘Result value.
Dim Services() As ENUM_SERVICE_STATUS ‘A list of the service status values.
Dim BufferBytesNeeded As Long ‘The size buffer needed to hold the
data.
Dim ServicesReturned As Long ‘Number of entries returned.
Dim ResumeHandle As Long ‘Handle for next service entry.
Dim BufferElements As Long ‘Number of buffer elements required.
Dim Outstring As String ‘Final output.
Dim ResultString As String ‘Intermediate output from array.
Dim counter As Long ‘Loop counter.

‘Open the Service Control Manager (SCM).

SCM_Handle = OpenSCManager(vbNullString, _
SERVICES_ACTIVE_DATABASE, _
SC_MANAGER_ENUMERATE_SERVICE)
If SCM_Handle = 0 Then
MsgBox “Failed to open SCM. DLL Error is: “ _
+ CStr(Err.LastDllError), _
vbOKOnly Or vbCritical, _
“SCM Open Error”

Exit Function
End If

‘Determine the buffer size required. We should recieve an error

‘stating the buffer isn’t big enough. The returned values will
‘contain the required buffer size.
BufferSize = 0
Result = EnumServicesStatus(SCM_Handle, _
SERVICE_WIN32, _
SERVICE_STATE_ALL, _
&H0, _
BufferSize, _
BufferBytesNeeded, _
ServicesReturned, _
ResumeHandle)
If Not Err.LastDllError = ERROR_MORE_DATA Then
MsgBox “Couldn’t determine the required buffer size. DLL Error is: “ _
+ CStr(Err.LastDllError), _
vbOKOnly Or vbCritical, _
“SCM Open Error”

Exit Function
End If

‘Calculate the buffer parameters. Determine the number of elements

‘required. Redimension the array to contain that number of elements,
‘and then calculate the actual buffer size.
BufferElements = BufferBytesNeeded / Len(Services(0)) + 1
ReDim Services(BufferElements - 1)
 Writing a Server Status Viewer 175

Listing 6.1 Continued

BufferSize = BufferElements * Len(Services(0))

‘Retrieve the service status information.

ResumeHandle = 0
Result = EnumServicesStatus(SCM_Handle, _
SERVICE_WIN32, _
SERVICE_STATE_ALL, _
Services(0), _
BufferSize, _
BufferBytesNeeded, _
ServicesReturned, _
ResumeHandle)
If Result = 0 Then
MsgBox “Failed to retrieve service status information. DLL Error is: “ _
+ CStr(Err.LastDllError), _
vbOKOnly Or vbCritical, _
“SCM Open Error”

Exit Function
End If

‘Interpret the data.

For counter = 0 To ServicesReturned - 1
ResultString = Space(250)
Result = lstrcpy(ByVal ResultString, ByVal Services(counter).DisplayName)
ResultString = Trim(ResultString)
If Len(ResultString) > 1 Then
ResultString = Left(ResultString, Len(ResultString) - 1)
Else
ResultString = “N/A”
End If
Outstring = Outstring + ResultString + “ (“

ResultString = Space(250)
Result = lstrcpy(ByVal ResultString, ByVal Services(counter).ServiceName)
ResultString = Trim(ResultString)
If Len(ResultString) > 1 Then
ResultString = Left(ResultString, Len(ResultString) - 1)
Else
ResultString = “N/A”
End If
Outstring = Outstring + ResultString + “) “ Ch

ResultString = Space(250) 6
Select Case Services(counter).ServiceStatus.CurrentState
Case SERVICE_STOPPED
ResultString = “Service is Stopped”
Case SERVICE_START_PENDING
ResultString = “Service is Starting”
Case SERVICE_STOP_PENDING
ResultString = “Service is Stopping”
Case SERVICE_RUNNING
ResultString = “Service is Running”
Case SERVICE_CONTINUE_PENDING
ResultString = “Service is Going to Continue”
176 Chapter 6 Creating Remote Access Utilities

Listing 6.1 Continued

Case SERVICE_PAUSE_PENDING
ResultString = “Service is Going to Pause”
Case SERVICE_PAUSED
ResultString = “Service is Paused”
Case Else
ResultString = “Status Unknown or Error”
End Select
Outstring = Outstring + ResultString + vbCrLf
Next
GetStatus = Outstring

‘Close the SCM handle.

CloseServiceHandle (SCM_Handle)

End Function

As you can see, the code begins by opening a connection to the service control manager
(SCM) using the OpenSCManager() API function. The SCM is the central authority for
manipulating services on a Windows NT/2000/XP machine. As part of the request for
access, you need to specify the name of a database to open and the level of access required
for the call. We’re using the active database in this example—there’s also an inactive data-
base. The call also tells the SCM that we want to enumerate services. You can request many
other activities, such as changing the status of a service or adding a new one.
The code calls the EnumServicesStatus() function twice. The first call determines the size
of the buffer required to hold the service status information. This information doesn’t
directly correlate to the number of array entries used to hold the status information, so
the next step is to perform some calculations and redimension the Services array. We
then need to set the final size of the buffer so that we can pass it along on the second
EnumServicesStatus() call. The second call retrieves the service status information. Note
that the Services array is 0 based because this is an API call. Figure 6.3 shows the complex
data structure returned for each service.

Figure 6.2
The Services array is a
composite of two data
structures.

Notice that each array element consists of two data structures. The first contains the
enumerated service description, while the second contains the status information. The
DisplayName and ServiceName entries are actually pointers to strings. The component uses
 Writing a Server Status Viewer 177

the lstrcpy() API call to retrieve the actual string values. However, this call relies on a
pointer to a string buffer, not a Visual Basic string. ResultString contains a series of zeroes
to mark the end of the string. Unfortunately, the zeroes also make it impossible to concate-
nate the string properly, so we need to perform some adjustments as shown in the code.
The status information is a series of flags that we can detect using constants. Listing 6.1
shows how to detect the current service status. You’d use similar techniques to determine
other status information.
The component code ends with a call to CloseServiceHandle(). Make sure you deallocate
the handle or your component will quickly develop a memory leak.

Tips for Working with Server Status Information

Many utility applications revolve around maintaining the server in some way. At least some of
these utilities provide information alone; they don’t allow you to change the server configuration.
A network administrator needs this constant flow of information to maintain the network prop-
erly. Consider how many command-line utilities check basic information such as the server’s IP
address and the number of required hops to transfer data from one node to another. Microsoft
recognizes the need for information by the network administrator and constantly improves their
management suite, but there’s always room for more high-quality management tools.
Most of the services that you’ll create require some type of enumeration. The sample in this
section is a good example of such a utility. Servers often manage multiple copies of a single
object. Providing resources to multiple clients is at the core of the server, so it’s not too sur-
prising that you’ll enumerate (list) items regularly. Everything from the computers attached
to the server to the processes requiring service comes in multiples. Building good enumera-
tion routines and recognizing the need to enumerate data is an essential part of developing
useful server status utilities.
Obtaining service status consumes quite a bit of space, even if you ask for the least possible
amount of information. If you look at the Services MMC snap-in (Figure 6.3) you’ll see that
every Windows 2000 machine contains a number of services. Returning just the names and
status will require more space than the typical dialog box provides. That’s why you’ll nor-
mally use a text box or other control with scrolling capability to display the information.
SOAP seems to play an interesting trick with the data you transfer using it. When you send Ch
data formatted with carriage returns and linefeeds, the data arrives at the other end without
linefeeds. According to several vendors, this is the anticipated action of the XML parser.
6
The result is that the data that looks great in a dialog box won’t break within a text box.
Figure 6.4 shows an example of what you’ll see.
You’ll need to get around this problem in some way. The two best methods are character sub-
stitution at the client end or using an array to transfer the data elements separately. The sam-
ple application in this section of the chapter uses character substitution. Transferring a single
string is still more efficient than using an array. In addition, the character substitution process
isn’t difficult to implement. However, it’s a detail to consider as you create your application.
Trying to figure out some problems like this can become brain-wracking sessions unless you’re
already aware of the behavior of the components that transfer data when using SOAP.
178 Chapter 6 Creating Remote Access Utilities

Figure 6.3
Windows 2000 sup-
ports a number of ser-
vices—too many to list
in a dialog box.

Figure 6.4
The XML parser
always strips linefeeds
from outgoing data.

Don’t get the idea that you’ll always use a single string. Using an array to transfer formatted
data has advantages as well. For example, the sample application could display the data using
a grid similar to the one used in the Services MMC snap-in. I chose to create a simple status
display for the example, but you might want to provide more details, and using a grid is def-
initely beneficial. In short, strings are better for performance, while arrays provide more
flexibility in formatting the data at the client end.
 Writing a Server Status Viewer 179

The utilities you create must be efficient. After all, querying the server status isn’t the central
purpose of a server. Every resource your utility uses consumes resources that could service a
user need. Normally it’s a good idea to limit the scope of information your utility will pro-
vide. For example, you’ll notice that the sample application queries only the active database.
We could have limited the scope of the component in other ways to reduce the amount of
resources required. For example, we could have retrieved a list of only those services that
were running.
Given the limitations of SOAP, you’ll want to avoid some performance-enhancing tech-
niques. For example, it might be tempting to provide a partial list of data in the hope that
the partial list will satisfy the network administrator’s requirements. This technique works
when a live connection exists between the client and server. It doesn’t work well with one-
way connections because the client application must create a new session if the network
administrator requires more information. When working with a protocol like SOAP, you
need to provide a complete answer so the component answers the request in one trip.

Working with the ISAPI Listener

The process of working with the ISAPI Listener is about the same as working with the ASP
file. In fact, the coding change required is so small that you might not even notice. When
you want to work with an ASP file, you initialize the SOAP client like this:
<service name=”ServerState”>
<port name=”ServiceCheckSoapPort”
binding=”wsdlns:ServiceCheckSoapBinding”>
<soap:address
location=”http://winserver/soapexamples/ServiceStatus/ASPServerState.ASP” />
</port>
</service>

Note that this code appears in the WSDL file, not in your program code. The program code
will remain the same no matter which access method you use. Here’s the ISAPI version of the
same component. Notice that only a single line of code changes between the two implementa-
tions.
<service name=”ServerState”>
<port name=”ServiceCheckSoapPort”
binding=”wsdlns:ServiceCheckSoapBinding”>
<soap:address Ch
location=”http://winserver/soapexamples/ServiceStatus/ServerState.WSDL” />
</port> 6
</service>

The example program contains buttons for both versions of the call. You won’t notice a dif-
ference with this example on a test network. However, using the ISAPI form of the call can
save considerable time and resources. Microsoft suggests that you normally use the ISAPI
method of creating SOAP calls. Given that you don’t need to make changes to your client
code and that you can generate the required WSDL files using the WSDL Generator utility
supplied with the Microsoft SOAP Toolkit, experimenting with both forms is easy.
180 Chapter 6 Creating Remote Access Utilities

You might run into an odd problem when working with an ISAPI listener that you won’t
run into when using an ASP listener for the same application. The Internet Information
Server (IIS) Administrator for some versions of Windows has a bug that won’t allow
you to enter paths with spaces in them. This prevents you from working with the
SOAPISAP.DLL and might prevent your application from working. The first way to fix
this problem is to move the SOAPISAP.DLL to a new location without spaces in the
pathname. You can also use the Microsoft recommended technique of using the short
default pathname of C:\PROGRA~\COMMON~1\MSSOAP\BINARIES\SOAPISAP.DLL.
In addition, the Microsoft SOAP Toolkit has an error as of this writing. It says the
SOAPISAP file is located in the Binaries directory of the toolkit. The file is actually
located in the \Program Files\Common Files\MSSoap\Binaries directory. Make sure you
keep this in mind when looking for files to update.

The ASP method does allow more server-side processing before making the SOAP method
call. This particular feature makes the ASP method more flexible than using ISAPI. It’s also
the reason you’ll use the ASP method instead of the ISAPI method more often than not.
SOAP is still in its infancy, which means that server-side processing for the sake of interop-
erability isn’t out of the realm of possibility. After SOAP becomes an established standard,
however, using the ISAPI method will become prevalent, at least for simple exchanges that
don’t involve complex data or other advanced processing requirements.

Some developers are concerned about the potential security risks of exposing the
WSML file to the user when using an ISAPI listener. The WSML file must appear in the
same directory as the listener, which means you can’t place it in a secure directory in a
separate location. The user could potentially read this file and use its contents to com-
promise the application in some way, depending on the application component. IIS
must load the WSML file to enable access to the component. However, it loads the
component on behalf of the user, not for the user. This means you could remove read
access from the WSML file, which prevents the user from retrieving the file over the
Internet. The application will still work because IIS loads the component on behalf of
the user using the server’s access rights. This little change in security could keep details
of your application from prying eyes.

One piece of magic isn’t immediately apparent when working with the ISAPI method. How
does IIS know what to do based on the contents of the WSDL file alone? The fact is that
the WSDL file doesn’t provide enough information to make the application work. Open the
Home Directory tab of the Default Web Site Properties dialog box. Click Configuration
and you’ll see an Application Configuration dialog box like the one shown in Figure 6.5.
Notice that the highlighted entry is for WSDL files and that it points to the
SOAPISAP.DLL we’ve been talking about throughout this section. IIS creates an applica-
tion mapping that automatically calls upon the SOAPISAP.DLL to process WSDL files as
needed. In fact, this entry is one of the first places you should look if you’re having trouble
getting the ISAPI method to work.
 Writing a Server Status Viewer 181

Figure 6.5
The ISAPI method
relies on application
mappings to do its
work.

Figure 6.5 has another notable feature. Notice the Cache ISAPI Applications check box. IIS
checks this option by default because caching ISAPI applications improves system perfor-
mance. Unfortunately, caching the applications also causes problems when you’re trying to
work with DLLs that IIS accesses. Because the DLLs are in memory, you can’t do anything
with them after a call. Clearing this option allows you to debug your applications faster.
Make sure you check this option again before you begin performance testing. Otherwise, the
performance statistics you gather won’t reflect the realities of working with the application.

Creating the Client

The client application for this example is similar to the clients we’ve created in the past.
The two unique features of this client are that you can use multiple methods to call
GetStatus() and the client includes a little extra code to add the line breaks the XML parser
removes back in. Listing 6.2 shows one of the three methods for obtaining and displaying
the server status information.

Ch
Listing 6.2 cmdRemote_Click() Source Code
Private Sub cmdRemote_Click()
6
Dim Client As SoapClient ‘Create the SOAP client.
Dim Outstring As String ‘String used to hold the output data.

‘Set up an error handler.

On Error GoTo ErrorHandler

‘Create the connection.

Set Client = New SoapClient
Client.mssoapinit _
“http://winserver/soapexamples/ServiceStatus/ServerState.WSDL”, _
“ServerState”, _
“ServiceCheckSoapPort”
182 Chapter 6 Creating Remote Access Utilities

Listing 6.2 Continued

‘Get the Status.
Outstring = Client.GetStatus

‘The XML parser strips all carriage returns, so we

‘need to replace them in the string for proper output.
Outstring = Replace(Outstring, vbLf, vbCrLf)

‘Display the data.

txtOutput.Text = Outstring

Exit Sub

‘Display a message when an error occurs.

ErrorHandler:

MsgBox Client.faultstring, vbExclamation

End Sub

As you can see, this method relies on the same basic techniques we’ve used in the past. Notice
the use of the Replace() function to add the carriage returns back into the string. Of course,
you also have to remove the existing linefeed. Figure 6.6 shows the final utility output.

Figure 6.6
The Services Status
Check Utility shows
the current state of the
services on any
Windows NT/2000/XP
server.

Creating a Simple Employee Check-In Application

As corporations create more online connections for employees to use, they also want to
account for employee time better. It’s all part of new resource management techniques
designed to make companies more competitive. Of course, e-mail provides one method of
 Creating a Simple Employee Check-In Application 183

remote monitoring, but there are some cases where an employee has no correspondence.
One of the purposes of this application is to provide a simple means for an employee to log
in—to provide some indicator that he is still alive and working on company projects.

Using Schemas as Design Tools

Many people view WSDL and other schema definition file types as a one-way conversion. In other words, you
create a component and then use a tool to create a WSDL file based on that component. However, you can
also use XML schemas as design tools, which means you’d design the schema first and then create a compo-
nent that matches the schema.
Don’t confuse this idea with the need to design interfaces for your application first. During the design phase, best
practice dictates that you define all of the required interfaces between components so there’s no confusion as to
what a component should do. The use of schema-based programming takes the idea a step further by saying the
programmer will create the schema code (the WSDL file) before creating the component it will support.
This system would actually work well on a large-scale development. The designer could create the various
schemas and then ask subordinates to create the component that would match the schema and provide the
required output given a specific set of inputs. The use of the schema would give the subordinate flexibility in creat-
ing the component, yet would assure that all of the components within the application would work together based
on their various schemas. The existence of a schema file would also allow for easier component validation.
Obviously, you’d need a WSDL design tool to actually make this work. The WSDL tool would have to present
the XML it contains in some graphical format, just as other application design tools do today. Some tools
approximate that behavior now, but none of them provides a full graphical implementation that would allow a
designer to create an entire application. However, such tools are quite likely to appear on the market as SOAP
and its associated technologies mature.
Now you need to ask yourself a question. Which should come first, the schema or the component? The choice
depends on how you want to develop the application and your personal programming preferences. For some
people, the schema will come first because it allows a developer to see the big picture before delving into
application coding.

Unlike many of the applications we’ve looked at so far, this is a classic one-way information
example. The employee isn’t looking for feedback from the company. All the employee
needs to do is check in. This means a one-way communication from the client to the server
is all that we need to accomplish the task.
It might seem at first that SOAP would be perfect at performing this task because it’s a one-
way communication protocol. However, as we’ll see in the sections that follow, one-way Ch
communication remains elusive because of the assumptions made by toolkit vendors. We’ll
also talk about some ways to fix the communication problems in a way that will make your 6
application run more efficiently.

Creating the Component

Active Directory opens many doors for the developer. As Windows 2000 becomes more
prevalent, developers will use Active Directory as a corporate resource database. You could
use this database to store anything. Theoretically, the database is completely extensible,
which allows you to add entries as needed. The hierarchical format of the database allows a
lot of flexibility in the way you define schemas for your data.
184 Chapter 6 Creating Remote Access Utilities

This example looks at Active Directory as a means for keeping track of employees on the
road. The section isn’t a full primer on Active Directory—that would require another book,
but I do include enough information to understand the example. Obviously, you could track
the employees in other ways, but this technique has the advantage of allowing employees to
check in as time and connections permit.
The Active Directory Services Interface (ADSI) allows you to work with Active Directory
using a standard set of interfaces. We’re using one of those interfaces in this example.
Listing 6.3 shows one of two methods used to create an Active Directory user object. This
object allows access to the Notes field of the Telephones tab (see Figure 6.7) of the User
Properties dialog box found in the Active Directory Users and Computers MMC snap-in.

Figure 6.7
The QCOne compo-
nent writes entries to
the Notes field on the
Telephone tab of the
User Properties dialog
box.

Listing 6.3 Quick Check-In One Component Source Code

Option Explicit

‘Define some AD constants.

Const ADS_PROPERTY_CLEAR = 1
Const ADS_PROPERTY_UPDATE = 2
Const ADS_PROPERTY_APPEND = 3
Const ADS_PROPERTY_DELETE = 4
Const E_ADS_PROPERTY_NOT_FOUND = &H8000500D

Public Sub CreateLogEntry(ByVal strUserName As String)

‘Create the variables required for the log entry.
Dim oUser As IADsUser
Dim strLDAP As String
Dim strLog As String

On Error GoTo ErrorHandler

 Creating a Simple Employee Check-In Application 185

Listing 6.3 Continued

‘Build a connection string.
strLDAP = “LDAP://winserver/CN=” + _
strUserName + _
“,CN=Users,DC=DataCon,DC=com”

‘Get the user object from Active Directory.

Set oUser = GetObject(strLDAP)
oUser.GetInfo

‘Create a new log entry string.

strLog = oUser.Get(“info”) + _
vbCrLf + “User Logged In: “ + _
Date$ + “ “ + Time$

‘Enter the string in the Notes field.

oUser.Put “info”, strLog
oUser.SetInfo

‘Release the user object.

Set oUser = Nothing

Exit Sub

ErrorHandler:
If Err.Number = E_ADS_PROPERTY_NOT_FOUND Then
‘This is the first time the user has logged
‘in so we need to create a new entry.
strLog = “User Login Times” + _
vbCrLf + “User Logged In: “ + _
Date$ + “ “ + Time$

‘Enter the string in the Notes field.

oUser.Put “info”, strLog
oUser.SetInfo

‘Release the user object.

Set oUser = Nothing

Else
‘Make a log entry if an error occurs.
App.LogEvent “Error Number: “ + CStr(Hex(Err.Number)) + vbCrLf + _
“Error Description: “ + Err.Description + vbCrLf + _ Ch
“Error Source: “ + Err.Source, _
vbLogEventTypeError 6
End If
End Sub

Public Sub ClearLogEntries(ByVal strUserName As String)

‘Create the variables required for the log
‘deletion.
Dim oUser As IADsUser
Dim strLDAP As String

On Error GoTo ErrorHandler

186 Chapter 6 Creating Remote Access Utilities

Listing 6.3 Continued

‘Build a connection string.
strLDAP = “LDAP://winserver/CN=” + _
strUserName + _
“,CN=Users,DC=DataCon,DC=com”

‘Get the user object from Active Directory.

Set oUser = GetObject(strLDAP)
oUser.GetInfo

‘Delete the string from the Notes field.

oUser.PutEx ADS_PROPERTY_CLEAR, “info”, Hex(Null)
oUser.SetInfo

‘Log an event so the network administrator knows

‘the entries were cleared.
App.LogEvent strUserName + “ cleared all login events”, _
vbLogEventTypeInformation

‘Release the user object.

Set oUser = Nothing

Exit Sub

ErrorHandler:
‘Make a log entry if an error occurs.
App.LogEvent “Error Number: “ + CStr(Hex(Err.Number)) + vbCrLf + _
“Error Description: “ + Err.Description + vbCrLf + _
“Error Source: “ + Err.Source, _
vbLogEventTypeError
End Sub

One of the more difficult tasks is building an LDAP string that we can use to access a par-
ticular resource, but using the ADSI Edit utility provided with the Windows 2000 Support
Tools makes that task much easier. Figure 6.8 shows a hierarchical view of the domain I’m
working with for this example. Note that you can find any Active Directory object and view
its properties.
Right-click a user name within the hierarchical list and select Properties. You’ll see a
CN=User Properties dialog box like the one shown in Figure 6.9. Notice the Path entry at
the top. It might look like you can do anything with this entry, but ADSI Edit allows you to
select the entry and copy it by pressing Ctrl+C. You can paste the resulting string into your
code and know for sure that the LDAP string is going to be correct.

While we’re in the CN=User Properties dialog box, notice the other fields. You can use
this dialog box to explore the various user properties that Active Directory makes avail-
able. It’s one way to determine how to modify your applications to make best use of
Active Directory.
 Creating a Simple Employee Check-In Application 187

Figure 6.8
ADSI Edit allows you
to locate Active
Directory resources
quickly.

Figure 6.9
Getting the correct
LDAP path is easy if
you know where to
look.

Ch

You’ll instantiate oUser by using the GetObject() method. The GetObject() method
requires an object reference in the form of a string, which is where the LDAP string comes
into play. However, instantiating oUser isn’t enough to fill it with information we can use.
The code also calls on the GetInfo() method to populate the object with data.
Once we have an object with which to work, all we need to do is create a string to put into
it. This requires two steps. First, we’ll use the Put() method to add the data to the local
memory variable. Second, we’ll use SetInfo() to send the data to the server.
188 Chapter 6 Creating Remote Access Utilities

Notice the odd-looking error handler for this example. Active Directory treats all empty
(Null) values as not existing. So, when we try to use the Put() method on an empty info
(Notes) field, Active Directory reports an E_ADS_PROPERTY_NOT_FOUND error. The error han-
dler looks for this return error and creates a special entry for it. In short, this error allows us
to create a header for the Notes field.
This example provides more in the way of error handling than previous examples. During
testing, I found that the entire process of querying Active Directory is more error prone
than many other activities. A small change in the Active Directory structure can wreak
havoc with applications. Even a small error in the entry of the user name will result in an
error. Unfortunately, all the client will see is a generic error message stating the server
doesn’t support the requested action. Because the server likely supported this action yester-
day, the user will be understandably confused (and the network administrator might be
equally concerned). Maintaining good log entries is essential when working with Active
Directory using SOAP. In this case, the event log entry will contain the error number, the
description, and the source of the error. Note that the error number is in hexadecimal to
make it easier to match to known errors.
ADSI provides some extended methods that you’ll need to know about as well. The
ClearLogEntries() method looks much the same as the CreateLogEntry() method, but
it uses the PutEx() method as shown here to clear the entries from the Notes field:
‘Delete the string from the Notes field.
oUser.PutEx ADS_PROPERTY_CLEAR, “info”, Hex(Null)
oUser.SetInfo

Setting the field to a Null value alone won’t clear it. You need to specify the ADS_PROPERTY
_CLEAR constant that the PutEx() method provides to actually clear the field.

Some Caveats About WSDL Files

We’ve visited the topic of WSDL files frequently in the book, but it’s important to realize
the role they play in efficient SOAP communication for applications that support them. One
of the oddities of most WSDL generation utilities is that they always assume two-way com-
munication between the client and server. What happens, however, when you need only
one-way communication, as in this example? The component we’ve designed doesn’t pro-
vide a result value and the client isn’t expecting a return value.
Figure 6.10 shows the output of the WSDL Generator utility that comes with the Microsoft
SOAP Toolkit. I highlighted the request/response pair for the CreateLogEntry(). The
ClearLogEntries() method has a similar pair of entries. If you generate this WSDL file
using a product such as 4S4C, you’ll get a similar result. The WSDL file contains the
request/response pair even though the return trip is wasted.
One of the few WSDL generators that I found that will create a one-way communication
entry is psWSDL Wizard (see Appendix C, “Third-Party Tool Reference,” for details on
this product). Figure 6.11 shows a WSDL file generated using the product that allows
 Creating a Simple Employee Check-In Application 189

one-way communication. Creating a WSDL file that permits one-way communication

makes your application more efficient. In short, you’ll want to use a third-party product to
create the WSDL file for some utility applications, rather than rely on the native capability
of the toolkit installed on your system.

Figure 6.10
Most WSDL generators
create a request/
response pair even
when you don’t need it.

Figure 6.11
The psWSDL Wizard
helps you create com-
plex WSDL files that
improve application
efficiency.

Ch

6
190 Chapter 6 Creating Remote Access Utilities

You’ll find a comparison of three WSDL generator outputs for this example in the Chapter
06\Active Directory Component directory of the source code available from the Que Web
site. You can find the source code at http://www.quepublishing.com. It’s helpful to look at
the output provided by different generators because of coding and efficiency concerns for
SOAP applications. Of course, there are interoperability concerns to consider as well. You
might find that a WSDL generator that normally provides good interoperability fails in a
specific situation. Developers will continue having interoperability problems between plat-
forms until the SOAP specification is complete.

Creating the Client

The client application looks much like all of the other high-level API clients we’ve created
to date. Listing 6.4 shows what this client looks like.

Listing 6.4 contains only the code for the cmdMakeEntry_Click() method. You’ll find
the full client source code in the \Chapter 06\Remote Log Test directory of source code
available from the Que Web site. You can find the source code at http://www.
quepublishing.com. The \Chapter 06\Local Log Test directory contains an alternative
client you can use to test the component before moving it to your server.

Listing 6.4 Remote Test Client Source Code

Private Sub cmdMakeEntry_Click()
‘Create the SOAP client.
Dim Client As SoapClient
Dim Result As String
Dim ErrorMessage As String

‘Set up an error handler.

On Error GoTo ErrorHandler

‘Create the connection.

Set Client = New SoapClient
Client.mssoapinit _
“http://winserver/soapexamples/UserLog/QCOne.WSDL”, _
“QCOne”, _
“UserLogSoapPort”

‘Make the log entry and report success.

Client.CreateLogEntry txtUserName.Text
MsgBox “Thanks for checking in!”, _
vbOKOnly Or vbInformation, _
“Check In Success”
Exit Sub

‘Display a message when an error occurs.

ErrorHandler:

ErrorMessage = “Fault Code: “ + Client.faultcode + vbCrLf + vbCrLf + _

“Fault String: “ + Client.faultstring + vbCrLf + vbCrLf + _
 Creating a Simple Employee Check-In Application 191

Listing 6.4 Continued

“Fault Actor: “ + Client.faultactor + vbCrLf + vbCrLf + _
“Detail: “ + Client.detail
MsgBox ErrorMessage, vbExclamation
End Sub

Notice that the client call doesn’t expect a result in this case because the server won’t pro-
vide one. Using one-way calls saves time and resources. However, because the server doesn’t
provide feedback, we’ll still need to display something for the user. In this case, a simple
message suffices.

Testing the Application

Normally, you can look at a utility program and quickly understand what the output is going
to be. For example, the server information utility at the beginning of this chapter is rela-
tively easy to figure out. You can guess that it will output a list of services and their current
state.
The utility in this section is a little different. It would be easy to extend this utility to per-
form all kinds of tasks in the background. All the user would need to do is click a single but-
ton to start the process. This lack of two-way communication means that you have to test
such an application thoroughly. In addition, you’ll need to spend more time looking at
potential problems.
Let’s look at the output of the application first. After you start the application, click Make
Entry twice. Open the User Properties dialog box for the user you selected and you should
see a display similar to the one shown in Figure 6.12. Likewise, you’ll want to verify that the
Clear Entry button removes the entries shown in Figure 6.12.

Figure 6.12
The component will
normally provide user
login time entries in
the Notes field.

Ch

6
192 Chapter 6 Creating Remote Access Utilities

Let’s talk about the Make Entry button a bit more and answer the question of why you
pressed the button twice. The server-side component went through two different processes
to create these entries. Remember that the first time the application generated an error and
handled it by creating the primary login entry. The second call didn’t generate an error
because the Notes field had an entry in it.
Next, you’ll want to enter an erroneous value for the user name. You should see two event
log entries on the server. The first will state that the Active Server Pages service has
stopped. The second will show the error entry from the component. Figure 6.13 shows a
typical server-side error event log entry.

Figure 6.13
Server-side event log
entries will contain the
information needed to
troubleshoot the appli-
cation.

The current version of the Microsoft SOAP Toolkit has an interesting problem. If you use
the ASP method of accessing the component, you’ll see the event log entries as anticipated.
On the other hand, using the ISAPI access method normally prevents the component from
making event log entries. Microsoft is aware of the problem and will probably fix it before
you read this book. However, it’s important to check for this issue so that you don’t rely on
event log entries that will never appear.
The sample component also makes an informational event log entry every time a user
clears the log. It’s important to let the network administrator know when users are per-
forming tasks that might not meet with company guidelines. You’d want to use this tech-
nique when company policy allows a user to perform a task under certain circumstances,
but not under others.
The client-side event log entry isn’t as informative as the one on the server, but it does
appear every time an error occurs, no matter which access method you use. Figure 6.14
shows a typical example of a client-side event log entry. As you can see, it tells you that an
error occurred and nothing more.
 Project 193

Figure 6.14
Client-side event log
entries are only useful
in providing you with
confirmation that an
error occurred.

Project
We’ve looked at two types of utility programs in this chapter. The first requests information
from the server, while the second sends information to the server. Both utilities are small
and perform a single task well. The following sections look at a few projects you might want
to try based on the content of this chapter.

Creating Your Own Utility

Many developers find that utility programs are the best way to learn new techniques because
they’re small and easy to build. You can also target utilities at learning how to perform spe-
cific kinds of tasks. For example, I created an entire series of small utility programs when
learning to work with Active Directory. The small programs allowed me to concentrate on
one area at a time without getting overwhelmed.
Vendors created SOAP to overcome many obstacles, one of which is the steep learning curve
of other technologies. However, even SOAP takes a while to learn. Creating your own utility Ch
could be the way to learn how to perform certain types of tasks, such as dealing with user-
defined types (an issue we discuss in Chapter 7, “Creating Data Entry Forms and Surveys”). 6
Try creating one or two small utilities of your own that reflect the tasks for which you’ll use
SOAP in your business. Here are some steps you can follow to develop these utilities quickly:
1. Determine which type of utility you want to create.
2. Decide whether the utility will actually use the features provided by SOAP. Make sure
that SOAP isn’t overkill for the situation.
3. Define the interfaces for your utility. Look for ways to use simple types to transfer data.
Also check for potential data translation problems, such as the XML parser’s stripping
of linefeeds from the SOAP character stream.
194 Chapter 6 Creating Remote Access Utilities

4. Simplify complex problems so the utility you create will perform a single task well.
5. Create the component for your utility. Try various WSDL generators to see which one
performs most efficiently. Look for obvious interoperability problems.
6. Code the client-side utility and test. Look at data output, event log entries, and other
clues that your application works as anticipated.

Upgrading the Server State and User Check-In Utilities

I kept both utilities in this chapter short to make the SOAP concepts I was trying to demonstrate
easier to see. Both utilities do provide useful features and you might find that you want to expand
their capability. This section contains some of the ideas that I’m going to try as I have time.
The server state utility is handy if you want to check the state of the services on several dif-
ferent servers. The reason this utility is useful is that it’s more flexible and runs faster than
the MMC snap-in. You could add enhancements, such as the ability to enter server names or
to output the data to a file in a batch process. Because my test system has only one Windows
2000 server, I hardwired the utility for now. However, I’ll probably add the ability to choose
a server later because my system does have other Windows servers and workstations on it.
The current version of the utility shows only one piece of status information. Another way to
upgrade the server state utility is to display more status information. You could even output all
of this information to a text file and automatically parse it for trouble spots on your system.
The user check-in utility is an idea in the rough. The idea isn’t new, but the technology
behind it is. Any network administrator will tell you that automation for the road warriors
isn’t just nice, it’s a must if you want them to actually get some work done while on the road.
Of course, the best automation is the type that works automatically. The user check-in util-
ity could update the user’s status information in Active Directory, upload hours worked,
orders, or anything else your company needs to perform.
One of the issues that SOAP developers are coping with now is asynchronous communication.
This application is a perfect example of something that would benefit from asynchronous com-
munication. A user could upload information to a queue on the server every time he checks in.
Because the user isn’t expecting a reply, the server can work on the input as time permits.
Hopefully, I’ve given you some ideas on how to upgrade the two utilities in this chapter.
Here are some steps to follow when adding to the utilities.
1. Determine if the utilities will work as they are now. If not, decide what you’d like to see
added.
2. Create a new project and add the code from the book utilities to it. That way your base-
line code won’t disappear under a cloak of changes.
3. Add new features one piece at a time.
4. Perform local testing, then place the utility on the server and test with SOAP.
5. Contact me about your latest creation at [email protected]. I’m always interested in
hearing how people have used the code within my books. Knowing this information
helps me create better examples in the future.
 CHAPTER
7
Creating Data Entry Forms
and Surveys
In this chapter
Determining Which Data Entry Vehicle to Use 197
Shortcuts for Data Entry and Survey Applications 199
Creating a Simple Survey Form 208
Creating a Simple Data Entry Form Application 221
Security, Privacy, Performance, and Reliability Issues 227
Handling Data Entry and Survey Form Errors 232
Using Templates for Quick Forms 233
Using MIME for SOAP Applications 234
Project 236
196 Chapter 7 Creating Data Entry Forms and Surveys

For many readers, the title of this chapter probably evokes visions of the simple online
forms that businesses use to solicit customer opinions. Many Web sites today include places
for users to interact with a company using surveys or other forms for data entry. Companies
collect a vast amount of information using these online sites. In fact, the amount of informa-
tion collected using data entry forms is so large that many companies can’t imagine conduct-
ing business without them. Obviously, data forms and surveys for interacting with customers
are an important consideration.
Of course, customer input isn’t the only use for data forms and surveys. Much of the infor-
mation that companies collect is from sources that use a product or service that the com-
pany provides or at least supports. For example, today software vendors collect a lot of user
information with registration forms transferred over the Internet. SOAP represents a way of
making the process of collecting this information easier and less error prone.
However, even these two forms of feedback are only the tip of the iceberg. Employees can
also make support calls using online forms. Likewise, customers can request status informa-
tion and partners can assess service potential using forms. SOAP also represents one of the
best ways to support this two-way communication. This second form of communication is
so important that Microsoft and other vendors are giving it a special name—Web Services.
No matter what name you use or how you look at data entry forms and surveys, they still
amount to a formalized method for exchanging data. The form enforces a certain level of
continuity between requests and simplifies the software required to handle the data. In
short, this chapter looks at methods for organized data handling that also interacts with a
data storage technology, such as a database.
This chapter will introduce you to some of the techniques you can use to make data entry
forms easier for both the client and the server. These techniques also reduce development
time and the burden on the developer. I won’t try to cover techniques that you can find in
books on general programming, but will instead focus on techniques for online data han-
dling. Of course, I’ll provide you with links to sources of information that you might want
to use to get the general information not included with this chapter.
The first section of this chapter, “Determining Which Data Entry Vehicle to Use,” intro-
duces the concept of a data entry vehicle. It describes the forms of data entry and tells you
how well each form will work in a given situation. By the end of this section, you’ll decide
on a data entry form for your application and know why it’s the best choice.
The second section of the chapter, “Shortcuts for Data Entry and Survey Applications,”
provides you with some shortcuts you can use to make the development process a little eas-
ier. The formal nature of these applications means you can use some of the same techniques
to create most applications—the content is different, but the procedure is the same.
We’ll create two applications in the third (Creating a Simple Survey Form) and fourth
(Creating a Simple Data Entry Form Application) sections of the chapter. The first example
shows how to create a survey application. It provides one-way communication between
client and server. This example will contain the first Web page processing in the book. The
 Determining Which Data Entry Vehicle to Use 197

second example is a data entry application. This example shows you the mechanics of creat-
ing a SOAP message with attachments. This is just one of many ways to send complex data,
but it might provide the best choice when working with certain types of database manage-
ment applications.
The fifth section, “Security, Privacy, Performance, and Reliability Issues,” addresses the four
non-programming issues of critical importance to your data entry form or survey applica-
tion. Security is a growing concern for all types of data. Likewise, privacy issues have
received major press in the past few months. Ensuring you actually receive the data sent by
the user in good condition is another concern. Finally, we’ll address the performance hit
your server will receive when processing the data.
In the sixth section of the chapter, “Handling Data Entry and Survey Form Errors,” we’ll
discuss special error handling for form-based data. This error handling is in addition to the
other types of error handling we’ve already discussed in the book.
The seventh section of the chapter, “Using Templates for Quick Forms,” will talk about
using templates to create forms. This is especially important because companies tend to
change their forms often. Surveys run for a limited amount of time and vendors who use
surveys are always devising new surveys to test consumer reactions. Using templates for
form-based data is essential if you want to retain your sanity.
The eighth section of the chapter, “Using MIME for SOAP Applications” discusses how
MIME will eventually affect SOAP applications. Unfortunately, vendors haven’t imple-
mented MIME support yet, even though it’s part of the specification. This section tells you
what vendors plan to do in the future. It also tells how using MIME will change data entry
for the better.
Finally, the Project section will take you through some steps for creating your own data
entry project. This section helps you put everything you’ve learned in the chapter into prac-
tice. This Project section allows you to experiment with the material you’ve just learned by
working with a quick project.

Determining Which Data Entry Vehicle to Use

Entering information into the computer is a time-consuming, but necessary task. However,
some users end up entering the same data more than once because of the way vendors
develop applications today. Even if you can copy and paste data from one application to
another, you also have formatting and other issues to consider. In addition, application
upgrades often wreak havoc with data entry. For example, there was an instance when
Microsoft released a version of Word that was so incompatible with its predecessors that the
product required a patch almost upon release. Ch

You can view data entry as two elements: the data and the formatting (or the view of the 7
data). Data entry today is more a view of a particular piece of information than the data.
Consider the simple word processor. Sure, the words are important, but how the application
198 Chapter 7 Creating Data Entry Forms and Surveys

presents those words onscreen is even more important in the minds of users. Word processor
vendors provide all kinds of interesting ways to quickly format data and retain the formatting
from session to session. Yet, the data entry process is about the same today as it was when I
first started using computers many years ago.
XML can free applications from worrying about the format of the data. This is such a power-
ful concept that companies such as Sun Microsystems are currently testing it in the real world.
StarOffice is an open source productivity application that includes several common features
such as a word processor, spreadsheet, presentation manager, and database. The StarOffice dif-
ference is that the data is stored in XML. This means that unlike products such as Microsoft
Office, you should be able to move the data between platforms without problem.
Sun plans to offer StarOffice on several platforms, including Windows, Solaris, and Linux.
Sun eventually hopes to establish the StarOffice file format as a universal format for all pro-
ductivity applications, which means we might finally free ourselves of inter-application com-
patibility problems. As part of this effort, Sun has released the draft specification to
OpenOffice.org (http://xml.openoffice.org) for study by the open-source community.
We’ll see in the “Working with SQLXML” section of Chapter 8, “Providing Remote
Database Access,” that Microsoft is also working hard to move some of their data to XML
(or at least make it accessible in that form).
Of course, this begs the question of what this means to you as a developer. Choosing the
correct data entry vehicle is becoming less important than the means used to store and
transfer the data. Users are demanding more from their applications, and management
wants to be sure that the data users enter today will remain viable tomorrow. Therefore, the
question of which data entry vehicle to use is one of data storage and user ease of use.
Consider, for a moment, the new kinds of applications that XML is allowing developers to
create. Tim Bray, the co-inventor of XML, has launched a public Web site (http://map.net)
that shows how you can visually plot information on the Internet. This launch is part of his
new company, Antarti.ca. Tim calls the application Visual NET. It allows users to navigate
the Internet visually, rather than use the old system of URLs. It relies on a special database
setup that stores data in a common format yet allows the application to determine the means
of presentation. This is one of the first promising Application Service Provider applications
to appear on the market that breaks the old molds. Ultimately, protocols such as SOAP will
allow even broader application sharing—users might not know precisely where their applica-
tion executes in the future, and they won’t care because the data is there for them to use.

Vendors are creating new data storage and transfer standards based on XML on a daily
basis. If you don’t see the technology that you need today, it will probably be along
with the next vendor press release. The point is that while the world is embracing XML
and by extension SOAP in a big way, the proliferation of standards means we’ll eventu-
ally have too many to support. Look for the usual technology shakeout in a year or so—
the remaining standards are those that are truly useful.
 Shortcuts for Data Entry and Survey Applications 199

This chapter looks at two of many data entry vehicles: the data entry form and the survey.
Data entry forms typically provide a formatted means to enter data in a two-way relation-
ship with the server, while surveys often provide more freedom in a one-way relationship.
We’ll look at other data entry vehicles as the book progresses. It’s important to realize that
all of these applications ultimately rely on XML as a data transfer methodology, but could
rely on other techniques for data storage.
It’s often hard to point out the exact moment that a transition occurs because historically,
the people involved in the transition don’t understand what’s taking place. XML could
become the next big transition for data storage and SOAP might become the most popular
way of moving XML from one place to another. It’s clear, however, that we’re in some sort
of transition brought on by the connectivity that the Internet provides. As a developer, you
need to consider your data entry vehicle choices carefully. Not only do you need to think
about the user’s needs, but you also need to consider how the data is ultimately stored on
the server.

Shortcuts for Data Entry and Survey

Applications
Database applications, no matter how well designed, present challenges that you won’t find
in any other application type. The fact that a database application is responsible for moving
mission-critical data means there’s added complexity to consider. A database application has
to manage that data for many users at the same time. In short, database applications are the
most difficult to create and likely the type you’ll run into most often.
The following sections won’t solve all of your database application problems. I could write
entire books on just one database management system (DBMS), much less try to cover them
all in one section of a chapter. These tips will help you with some of the problems you’ll
experience working with databases under SOAP. We’ll discuss topics such as data conversion
and methods to reduce the number of round trips you have to make.

Reducing the Number of Round Trips

You’ve likely seen at least one database application where the developer uses several round
trips to accomplish any given task. For example, a session might begin with the application
requesting some bit of information to perform a lookup. After it performs the lookup, the
application will send part of the data back to the client and ask for still more input. The
client provides the additional input and the application sends back another part of the data
picture. This seesaw action continues until the query is complete.
Although the multiple-request approach might work on a LAN or WAN, it’s not going to Ch
work on the Internet. For one thing, SOAP is a one-way protocol that some vendors have
extended into a request/response format. Another problem is bandwidth. Making more than
7
one request is going to slow the application to the point that no one will want to use it.
200 Chapter 7 Creating Data Entry Forms and Surveys

Of course, this point actually comes down to a problem with design in most cases. You need
to design the application so that it performs well with a single request and a single response.
This means gathering all of the data needed to answer an entire query at one time. It might
also mean separating detail pages so that you get the main data on one page and the details
on another. Theoretically, this will allow some developers to get around the multiple request
problem without many application design changes.

Choosing Document or RPC Style WSDL Files

WSDL actually supports two formatting styles: document and remote procedure call (RPC).
The RPC style is the most common by far. In fact, it’s the only choice you’ll have when
using many WSDL generators. One of the few products that does give you a choice is the
psWSDL Wizard (see Appendix C, “Third-Party Tool Reference,” for details). Figure 7.1
shows you the dialog box that allows you to choose between document and RPC style.

Figure 7.1
Only a few WSDL
generators allow you
to choose between
document and RPC-
style output.

The RPC style is most popular because it allows the greatest flexibility in sending data
between client and server. You can’t use the document style to send abstract data types. This
makes the document style inappropriate for some database applications. Figure 7.2 shows
the RPC output for the ServerState.DLL example from Chapter 6, “Creating Remote
Access Utilities.” (You can also compare the two styles by looking at the WSDL files in the
\Chapter 07\Doc vs RPC directory of the source code available from the Que Web site.
You can find the source code at http://www.quepublishing.com.)
Now, compare Figure 7.2 to the document-style output in Figure 7.3. As you can see, the
document style is less complex and consumes less space. There are fewer opportunities for
interoperability problems when using the document style. In addition, the document style is
more efficient. It allows you to transfer data using fewer resources.
 Shortcuts for Data Entry and Survey Applications 201

Figure 7.2
The RPC style
output for the
ServerState.DLL
allows complex
data definitions.

Figure 7.3
The document style
output for the
ServerState.DLL
provides efficient
data transfer.

Ch

7
202 Chapter 7 Creating Data Entry Forms and Surveys

Don’t get the idea that the document style type is only for small amounts of data. Many
developers prefer to use the document style message for large amounts of simple data. For
example, a purchase order might contain only strings or a combination of strings and num-
bers. Transferring a large purchase order using the document style will consume fewer sys-
tem resources and enhance performance.
The choice of document versus RPC style is normally determined for you by the type of
data you want to transfer. Always use the RCP style when transferring abstract data or calls
that contain UDTs. However, in those situations when you do need to make choice, using
the document style might be the best choice because of the performance boost it provides.
The RPC and document styles differ in another way that will affect your decision. Most
developers view the RPC style as easily adapted to scripting. You would use it in situations
when you don’t know anything about the client in advance. The document style requires
advance knowledge of the document format. In short, many developers view the document
style as useful between business associates, but not between a business and multiple cus-
tomers. The SOAP specification doesn’t necessarily define this viewpoint; developers base
this opinion on experience with SOAP.
A final consideration is that at least some developers view the document style as easier to
document. The document style shows workflow more clearly, making the intended use eas-
ier to divine from the message alone. However, this is another situation where opinion and
personal experience might figure more highly than fact.

Empty and NULL Value Processing

One feature of database management programs is that not every field of every record will
contain data. Some records will contain empty or NULL values. For example, it’s common to
provide a second address line for client contact systems. Some clients will have a second
address line, some won’t; therefore, this field will contain an empty string in some cases.
Some developers will attempt to short-circuit message output for NULL and empty values by
providing a single tag output for the value. Unfortunately, the following output won’t work
in most cases:
<Address2>
<Value />
</Address2>

Most XML parsers will simply ignore the Address2 value, resulting in odd application
behavior. For example, some developers reported value shifting when they attempted to use
this technique. The XML parser would fill the missing value with the next value in line. You
can bypass this problem for empty values by providing an empty tag set like this:
<Address2>
<Value></Value>
</Address2>
 Shortcuts for Data Entry and Survey Applications 203

Of course, you might need a method for differentiating between empty and NULL values.
Unfortunately, the XML and SOAP specifications don’t provide a way to differentiate
between the two. However, many developers now use the word “NULL” in all uppercase, as
shown here, to differentiate between NULL and empty values.
<Address2>
<Value>”NULL”</Value>
</Address2>

Note that this entire issue is in a state of flux as I write this and a standardized solution
might be available when you read this section. The method used to define empty and NULL
values is important, and the lack of consistency is problematic. Some developers have
reported getting different results for empty and NULL values by changing the product ver-
sion. The point is that if you plan to transfer empty or NULL values, you should perform test-
ing on every platform you plan to support. Otherwise, you might end up with many
unexpected support calls that vary according to the product version in use.

Unregistering Your Control

Database management applications seem to require more than the usual number of controls.
Registering and unregistering these controls as you test them can become a painful experi-
ence because you need a DOS command prompt every time. I always keep a command
prompt handy for just this purpose.
You can make the process of creating a command prompt where you need it easier by
adding a new entry to an existing file type in Windows. Locate the Folder File Type on the
File Types tab of the Folder Options dialog box in Windows Explorer. Click Advanced.
You’ll see an Edit File Type dialog box. Click New. You’ll see a New Action dialog box.
Type Command Prompt Here in the Action field and cmd.exe /k “cd %1 in the Application
Used to Perform Action field. Click OK three times to complete the action. Now, every
time you need a command prompt, right-click the folder and choose Command Prompt
Here from the context menu.
Some people find that they don’t like using the command prompt to register and unregister
their components using RegSvr32. You’ll find an alternative method in the Xteq COM
Register Extension at http://www.xteq.com/downloads/index.html#comr. This utility adds
two new menu entries to your system as shown in Figure 7.4.
All you need to do is right-click the component and choose Unregister library to unregister
it. Likewise, choosing Register library will add it to the registry. You get the same RegSvr32
messages as normal. The only difference is that you don’t have to go to the command line to
perform the task.
Ch

7
204 Chapter 7 Creating Data Entry Forms and Surveys

Figure 7.4
The COM Register
Extension adds new
menu entries to
your system.

Understanding CDATA Sections

XML supports the concept of a CDATA (or character data) section within messages. You
can use a CDATA section to prevent interpretation of part of the message by the XML
parser. For example, this is one way to get around the problems with transferring carriage
return/linefeed pairs in a message. This sounds like a promising fix for many problems that
developers experience when using SOAP. Using a CDATA section is easy. All you need to
do is enclose the CDATA as shown here:
<![CDATA[My Data]]>

Unfortunately, CDATA sections aren’t the cure-all that developers think they are. Many
developers experience problems using CDATA sections because they don’t understand the
restrictions of using them. For example, you can’t place the ]]> character sequence any-
where in the section for the obvious reasons—the XML parser will think the CDATA sec-
tion has added and you’ll get unpredictable results.
This means that you can’t use a CDATA section to transfer an XML document because you
can’t be sure the XML document won’t include the ]]> character sequence. The CDATA
sections are limited to clean character data—data that is primarily text without control
information. Although you can transmit functions using CDATA sections, you must ensure
that the function won’t contain data that might cause problems within the CDATA section.
CDATA sections are a valuable asset for the developer. However, you’ll want to ensure that
you understand the limitations of using them. Attempting to use CDATA sections when you
need to use Base64 encryption or attachments will definitely cause crashes and unpredictable
application behavior.

Understanding XML Document Transmission Restrictions

Someone has said that the odd thing about SOAP is that it’s XML technology to the fullest,
but it can’t easily transmit an XML document. You read in the previous section that you
 Shortcuts for Data Entry and Survey Applications 205

can’t use CDATA sections to transmit XML documents because you can’t be sure the docu-
ment will be free of unwanted character sequences.
Not only do you have to consider character sequences in XML documents, but you also
have to consider its encoding. Your SOAP application might rely on UTF-8 for example,
whereas the XML document you want to transfer relies on UTF-16. The difference in
encoding will cause your application to crash or at least act unpredictably.
One way to overcome this problem is to transfer the XML document as an attachment.
Unfortunately, none of the presently available SOAP toolkits supports attachments. If you
want to use attachments, you have to build your own functionality from scratch, which is a
daunting task for the best programmer.
Another way to overcome the problem of transferring XML documents is to use Base64
encoding. This is the current method used to transfer binary data. It works well, but exacts a
performance penalty. Base64-encoded data requires more space; therefore, it requires more
time and resources to transfer. Using Base64 encoding on large documents might prove pro-
hibitive from a performance perspective. Some developers also worry about compatibility
problems when using Base64, although no one has reported major problems in this area.
Some developers get around the problem by using escape sequences in place of problem
markup characters. The document object model (DOM) parser at one end converts prob-
lematic character sequences to escape sequences, whereas the DOM parser on the other end
converts the escape sequences to markup language character sequences. This method still
exacts a performance penalty, but not as great as using Base64 encoding. However, using
this method restricts you to using the DOM parser. Many developers prefer using Simple
API for XML (SAX) parsers because they provide a performance boost. In addition, using
escape sequences means you must examine the code carefully to ensure it catches all of the
markup language characters.

Some XML parsers will automatically handle escape character sequences for you when
used in DOM mode. This saves the developer time coding what should be a standard
feature. However, you also need to verify that the XML parser provides consistent
escape character sequence handling. For example, some developers report that the
Microsoft XML parser (MSXML) version 3.0 handles the vertical tab character (ASCII dec-
imal 11) unpredictably. In some cases, it escapes the character; in other cases, it doesn’t.
Fortunately, the parser will register an error to tell you about the problem character. The
reason that I mention the vertical tab is that it’s an illegal character according to the
specification (http://www.w3.org/TR/2000/REC-xml-20001006#NT-Char), so
MSXML should always fail or at least strip the character out of the data stream. The
point is that you want to look for an XML parser that provides predictable behavior if
you plan to use it for adding escape character sequences automatically. Ch

7
The bottom line is that no perfect solutions to this problem exist. You need to choose the
solution that causes the least number of problems in your program environment, and then
find ways around the remaining problems. For example, you could performance tune your
application to reduce the effects of Base64 encoding performance penalties.
206 Chapter 7 Creating Data Entry Forms and Surveys

Using a Third-Party Product to Document Your Components

Of all the applications we’ll create in the book, it’s most likely that someone will need access
to your components when working with survey or data entry applications. A partner might
require access to these components to incorporate them into an application that accesses
resources on your site. For example, ordering parts might consist of creating a data entry
form that sends the information directly to your company.
Many products on the market help you build documentation for your components.
However, most are difficult to use and many are quite expensive. DocBuilder is a freeware
product created by GFaI. The terminology has a definite German slant, but you can easily
modify it for use with English by changing the configuration files. GFaI supplies well-writ-
ten English language documentation, and I found that modifying the scripts was relatively
easy using the tips in the frequently asked questions section of the help file.

We’ve discussed the importance of documenting your components in situations where a

third party might need to use them. DocBuilder (http://www.gfai.de/produkte/
docbuilder/e_index.htm) relies on scripts to search your application and create doc-
umentation automatically. VBdocman (http://vbdocman.exit.mytoday.de/), a
shareware documentation product, creates help files for your Visual Basic applications in
HTML, RTF, and Windows Help formats. You can also create user-defined output formats,
such as XML, with this product. As with DocBuilder, you don’t have to add comments
manually to use this product, but it helps. The quality of the documentation you receive is
directly proportional to the number of tags added to the documentation. Unlike some
products on the market, VBdocman will add the tags to the document for you, but you
still have to add descriptive text. The generated file can include see also links, example
links, objects, methods, property or event descriptions, parameter descriptions, setting
descriptions, return values, and remarks. You’ll find an example of the output for
VBdocman in the \Chapter 07\VBdocman directory of the source code available from the
Que Web site. You can find the source code at http://www.quepublishing.com.

The product in this section works only with C++, Pascal, and Delphi at the time of this
writing. However, the author, Dr. Josef Richardt, intends to add Visual Basic support in
the future. You can also add the specially formatted comments to a Visual Basic file and
still use DocBuilder with it; the product won’t work with an uncommented Visual Basic
file. Consequently, the example will show a Visual C++ example and not a Visual Basic
example. Testing shows that DocBuilder works equally well with Borland C++ and Delphi.

This section shows how to use DocBuilder with the ServInfo component in Appendix D,
“SOAP for Visual C++ Developers.” Using this product is easy. You’ll want to move one of
the script files to the project directory from the \Program Files\DocBuilder\examples\dbs.
Create a new project using the File | New Project command. Define the project and output
directories, along with the script file that you copied to the project directory. Figure 7.5
shows a typical New Project dialog box.
 Shortcuts for Data Entry and Survey Applications 207

Figure 7.5
Typical DocBuilder
New Project dialog
box containing
directory and
script information.

Notice the Define Input Files button at the bottom of the dialog box in Figure 7.5. Click
this button to add files to your project. You can add the files later, but it’s more convenient
to add them now. In this case, I’ll add the ViewServer files from the ServInfo component
because these files contain my class descriptions. Figure 7.6 shows how the Input Files dia-
log box looks after you add files to it. As you can see, DocBuilder allows you to create com-
plex setups consisting of multiple components. Click OK to close this dialog box, and click
OK again to close the New Project dialog box.

Figure 7.6
Use the Input Files
dialog box to add class
files to your project.

At this point, you’ll see the main project window. Most of the menus act as you’d expect.
You can use the Definitions menu to redefine elements of your project such as the script
used to create output. You’ll also use this menu to title your document and to modify help
file attributes. (I found the defaults worked fine for the components that I tested.)
After you have the project defined, use the options on the Analysis menu to check your
setup. You must perform an analysis before you can create output files. The analysis process
will show any errors in your project files that prevent DocBuilder from creating documenta- Ch
tion for your component. 7
208 Chapter 7 Creating Data Entry Forms and Surveys

The final step is creating output. You have a choice of HTML, Windows Help, or rich text
format (RTF) files. I used the Windows Help file format for the example. Figure 7.7 shows a
typical Windows Help table of contents for this product when working with a simple com-
ponent. Notice that you can obtain a full description of almost every element of the compo-
nent. Of course, this is a simple component, so there was little information to garner, but
tests on complex components show that the product scales well.

Figure 7.7
Windows Help
contents output
from DocBuilder.

Figure 7.8 shows the type of documentation you can expect using an unmodified code file.
Adding comments to your source code will improve the documentation produced. Notice
the German language title bar entry. You can change these strings to English or any other
language by modifying the output files. For example, you’d modify the help project (HPJ)
file for a Windows Help project.

Creating a Simple Survey Form

Companies use survey forms for a variety of purposes. The most common use is to ask cus-
tomers how they feel about a product and how they use it. In some cases, a company might
ask for customer referrals or provide an opportunity for the customer to ask for additional
information about another product. Survey forms also figure prominently in employee rela-
tions. The suggestion form is nothing more than a survey in disguise. From a development
perspective, many forms follow the survey format.
A good way to look at surveys is as data entry forms that require a one-way communication.
The form requests data in the form of an opinion and the sender knows there won’t be any
immediate response (except a simple thank you message at the end). Most survey forms
include two question types. The first requests a definite response, such as rating the usability
 Creating a Simple Survey Form 209

of a product for a specific purpose. The second is an essay answer where the respondent
types an answer. In short, you can reduce survey data to integers and strings, making it easy
to transfer the data using any method.

Figure 7.8
A sample of an
individual document
page from DocBuilder.

Some developers will work with more than one company, each with its own set of
components. This leads to using more than one WSDL file. A simple fact to remember
is that a SOAP client can handle only one WSDL file. You can combine multiple SOAP
clients in one application, but you must maintain separate clients for each WSDL file.

Surveys don’t provide much value if you don’t analyze the data they provide. Some compa-
nies place the information in a spreadsheet and perform “what if” analysis for days at a time.
This method works best with complex forms that request a lot of input of various types.
Other companies create an application to perform the required analysis and output the
results to anyone within the company who needs the information. This method works best
with simple forms where the company is expecting a specific type of information output.
The following sections show you how to create a simple survey form that stores its results in
a database. After you have the input application created, the sections that follow will show
how to create an output application. Your company will require both application types to use Ch
survey data.
7
210 Chapter 7 Creating Data Entry Forms and Surveys

Creating the Database

The database for this example is extremely simple. It contains a combination of integer and
string fields that track user survey input. The example won’t contain any odd data formats to
translate or unexpected surprises from a data translation perspective. (See Chapter 8 for
examples of data conversion problems.) Figure 7.9 shows the structure of this database.

Figure 7.9
A diagram of the
simple database used
in this example.

This example (and the one in the following section) relies on Microsoft SQL Server 7.0.
The scripts found in the \Chapter 07\Data directory of the source code available from
the Que Web site allow you to re-create the database within SQL Server. You can find
the source code at http://www.quepublishing.com The directory also contains a
backup of the data so that you don’t have to enter it manually.

As you can see, the survey will ask questions about the food likes and dislikes of the people
surveyed. The database uses the most efficient method possible to store these likes and dis-
likes in an anonymous manner. Given privacy concerns today, many companies will resort to
the anonymous survey when all they need is the information the respondent can provide.
Of course, we still need some way to identify the individual entries, and the survey form
doesn’t provide a means to guarantee uniqueness. The EntryNumber field contains a simple
number that uniquely identifies each record. SQL server will update this field automatically.
Notice that this is the only field used for the primary key.
This example won’t cover some obvious additions to this database, such as individual field
indexes to allow fast data searches. The only intent of this example is to show the SOAP
connection. You’d normally add filtering and indexing to reduce the data returned to the
user and the amount of processing time required.

Creating the Server-Side Component

The server-side component accepts input from the survey form and places it in the database.
Depending on the clients that you intend to use, the server-side component could also per-
form data analysis to ensure the input data meets specific value range and format criteria.
However, this example will perform the range checking on the client—the option you
should choose in most cases.
 Creating a Simple Survey Form 211

Performing value range, data type, and format checks on the client using scripts reduces
processing load in several ways. First, it allows the user to receive instant feedback on entries
as they’re made. This reduces user confusion and makes it more likely the data will appear
in the correct format the second time the user enters it. Second, the server doesn’t have to
perform the analysis. The server-side component can assume the data is in the correct for-
mat before it arrives. Finally, every verification step requires a call to the server if you use
server-side processing. Using client-side processing reduces application bandwidth require-
ments and speeds entry.
Now that you understand the coding basis for this component, let’s look at some code.
Listing 7.1 shows the component source code. Notice the component contains only the
input code. The output code (the thank you note) resides on the client in script form, as
we’ll see later.

Listing 7.1 Data Survey Input Component Source Code

Public Sub EnterFoodResults(ByVal LikeMeat As Integer, _
ByVal LikeVegetable As Integer, _
ByVal LikeFruit As Integer, _
ByVal LikeGrain As Integer, _
ByVal DislikeMeat As Integer, _
ByVal DislikeVegetable As Integer, _
ByVal DislikeFruit As Integer, _
ByVal DislikeGrain As Integer, _
ByVal Comments As String)

Dim Conn As ADODB.Connection ‘Database connection.

Dim Recc As ADODB.Recordset ‘Recordset
Dim ConnStr As String ‘Database connection string.

‘Create a connection string.

ConnStr = “Provider=sqloledb;” & _
“Data Source=WinServer;” & _
“Initial Catalog=Survey;” & _
“User Id=sa;Password=; “

‘Connect to the database.

Set Conn = New ADODB.Connection
Conn.Open ConnStr

‘Obtain the required recordset.

Set Recc = New ADODB.Recordset
Recc.Open “Food”, _
Conn, _
adOpenStatic, _
adLockOptimistic, _
adCmdTable
Ch
‘Enter the data.
Recc.AddNew 7
Recc!LikeMeat = LikeMeat
Recc!LikeVegetable = LikeVegetable
Recc!LikeFruit = LikeFruit
Recc!LikeGrain = LikeGrain
212 Chapter 7 Creating Data Entry Forms and Surveys

Listing 7.1 Continued

Recc!DislikeMeat = DislikeMeat
Recc!DislikeVegetable = DislikeVegetable
Recc!DislikeFruit = DislikeFruit
Recc!DislikeGrain = DislikeGrain
Recc!Comments = Comments
Recc.Update

‘Close the recordset and database.

Recc.Close
Conn.Close

‘Clear the variables.

Set Recc = Nothing
Set Conn = Nothing

End Sub

As you can see, this is essentially a data access component that adds new records. You don’t
need to provide any other capability within this component because the application won’t
allow the user to modify his input after he submits the survey. In fact, it might be counter-
productive to add other features because this would only open the door to unwanted cracker
activity—keeping the capability of the component limited is one form of security precaution.
The component uses the following five-step process to submit a survey:
1. Create a connection to the database.
2. Open the recordset. In this case, we’ll use a table because the code is simply adding
a record.
3. Create a new record.
4. Add content to the new record, and then update the record on the database. (The new
data won’t appear on the server unless you use the Update() method.)
5. Close the recordset and the connection to the database. If you fail to close the database
connections and free the object resources, the application will leak resources and you’ll
eventually need to reboot the server.

Notice that the example uses “sa” as the user ID and that there isn’t a password. Normally,
you’d use a custom user ID and password within the code. Make sure you keep the database
secure by providing an account that can only add records for your surveys. This code addi-
tion works with the component limitations to reduce your risk in allowing unknown third
parties to add survey information to a database.

Designing a Survey Form

You can use any of a number of survey form input types. The example uses a form-based
application, but you could easily use a Web page or other variations. The client is providing
input alone; the client doesn’t care about any data the server might return, so anything that
can generate a SOAP message as output will work fine in this case. Listing 7.2 shows the
client source code for this example.
 Creating a Simple Survey Form 213

Listing 7.2 Data Survey Input Client Source Code

Private Sub cmdQuit_Click()
‘Exit the application.
End
End Sub

Private Sub cmdSubmit_Click()

‘Create the SOAP client.
Dim FoodSurvey As SoapClient
Dim ErrorMessage As String

‘Set up an error handler.

On Error GoTo ErrorHandler

‘Create the connection.

Set FoodSurvey = New SoapClient
FoodSurvey.mssoapinit _
“http://winserver/soapexamples/Survey/Survey.WSDL”, _
“Survey”, _
“FoodSoapPort”

‘Enter the survey data.

FoodSurvey.EnterFoodResults LikeMeat.ListIndex, _
LikeVegetable.ListIndex, _
LikeFruit.ListIndex, _
LikeGrain.ListIndex, _
DislikeMeat.ListIndex, _
DislikeVegetable.ListIndex, _
DislikeFruit.ListIndex, _
DislikeGrain.ListIndex, _
Comments.Text

‘Release the object.

Set FoodSurvey = Nothing

‘Display a success message.

MsgBox “Thank you for completing the survey!”, _
vbInformation & vbOKOnly, _
“Survey Entered Successfully”

‘We’re finished, so exit.

Exit Sub

‘Display a message when an error occurs.

ErrorHandler:

ErrorMessage = “Fault Code: “ + FoodSurvey.faultcode + _

vbCrLf + vbCrLf + _
“Fault String: “ + FoodSurvey.faultstring + _
vbCrLf + vbCrLf + _
Ch
“Fault Actor: “ + FoodSurvey.faultactor + _
vbCrLf + vbCrLf + _
“Detail: “ + FoodSurvey.detail
7
MsgBox ErrorMessage, vbExclamation
End Sub
214 Chapter 7 Creating Data Entry Forms and Surveys

Listing 7.2 Continued

Private Sub Form_Load()
‘Set the combo boxes to known values.
LikeMeat.ListIndex = 0
LikeVegetable.ListIndex = 0
LikeFruit.ListIndex = 0
LikeGrain.ListIndex = 0
DislikeMeat.ListIndex = 0
DislikeVegetable.ListIndex = 0
DislikeFruit.ListIndex = 0
DislikeGrain.ListIndex = 0
End Sub

As you can see, the client creates a SOAP client, outputs the data, displays a success mes-
sage, and exits. Notice the success message isn’t based on server feedback—it only acknowl-
edges the user’s participation in the survey. It’s unlikely the user will participate in the survey
a second time even if the client fails to deliver the requested data, so acknowledging the
input is about all you can do.
Notice this example uses a detailed error message. The application could be located in a
kiosk or other remote location where the person monitoring the application has limited
skills. If the application fails, you want to create the most detailed feedback message possible
so the person reporting the problem doesn’t have to fiddle with the application much. In
fact, the error reporting for this type of application should contain as much detail as you can
provide. Of course, thorough testing will limit the number of times you actually need to rely
on this information to fix an errant application.

Testing the Input Application

As with any other application, it’s important to test this one thoroughly. Of course, one of
the first tasks you’ll perform is to move the component to the server, register it, and then
construct a WSDL file. It’s especially important to check the WSDL output in this case.
This is a survey form, so you don’t want the component to pass information back to the
client. This is a one-way trip as far as the client is concerned. Figure 7.10 shows the WSDL
file for this example. Notice there isn’t any feedback—not even a Result variable (as you’d
receive when using a function within Visual Basic). The <message name=”Food.
EnterFoodResultsResponse” /> entry signifies that this is an empty tag. Contrast this with
the Food.EnterFoodResults tag, which contains elements for all of the input values.
Figure 7.11 shows the main dialog box for this application. Notice that it uses combo boxes
that don’t allow any kind of input. This reduces the chance that the user could enter a
wrong value because he can only select from a list of predefined values. The test box is also
set to limit the user’s input to 100 characters. The user is inputting a comment, in this case,
so checking the content of the comment is unnecessary except for ensuring it isn’t too long.
 Creating a Simple Survey Form 215

Figure 7.10
Always verify that the
WSDL file provides
nothing in the way of
a return value when
working with surveys.

Figure 7.11
Careful application
design will limit the
amount of error
checking code
required in your
application.

These careful design choices are why the client code in Listing 7.2 lacks any form of input
error check. Depending on your application, you can significantly reduce or at least limit the
amount of input checking code you need to create. Using this technique reduces the chance
that the application will fail because there are fewer failure points to consider. Input check-
ing is essential for this kind of application because the user might not have much in the way
of computer skills.
Testing the client application is relatively simple. Add some values to the application, and
then click Submit. You should see the success message shown in Figure 7.12. Of course, this Ch
message only means that the client application works. It doesn’t signify that the data actually
arrived at the server or that the server-side component entered data into the database. The
7
lack of a SOAP failure message only means that SOAP sent the data to the server because
there’s no return value to test.
216 Chapter 7 Creating Data Entry Forms and Surveys

Figure 7.12
This simple success
message shows that
the client application
works.

As part of application testing, I normally design a dialog box-based application with a grid of
the data values. You’ll find this application in the \Chapter 07\Data Edit directory of the
source code available from the Que Web site. You can find the source code at http://www.
quepublishing.com Figure 7.13 shows what this application looks like. You’ll want to try all
of the values that the program can accept and then try various combinations. Nine or ten
successful tests normally indicate the application is ready for use with a simple application.
Complex applications, especially those with many fill-in-the-blank answers, require more
testing.

Figure 7.13
Create test applica-
tions that allow you to
view the results of
your survey client.

Make sure you erase any test data before you release the application to production
use. At least one developer recently left the test data in place by accident. The client
conducted the survey and began compiling the results. The list of accidentally included
test messages skewed the data and invalidated the survey results. The company was
able to recover by eliminating the test messages, but the developer lost a customer in
the process. In short, make sure your testing phase includes cleaning up afterward.

Writing an Analysis Component

Surveys are useless if you can’t analyze their content and create a report. We used SOAP for
the survey because the input application could appear anywhere. Using SOAP allows the
survey application to send data to a remote database. Because surveys are usually one-time
 Creating a Simple Survey Form 217

applications, many companies will use an outside consultant to create the survey and provide
analysis for the results. Therefore, it’s quite likely that you’ll need to use SOAP for the
analysis application as well. The survey components and results reside on your local server.
The client will want to access these results as often as needed, so using SOAP to transfer
data to a client application on their local workstation makes sense. Listing 7.3 shows the
analysis component for this example.

Listing 7.3 Analysis Component for Food Survey Source Code

Public Function GetFoodData() As Integer()
Dim Conn As ADODB.Connection ‘Database connection.
Dim Recc As ADODB.Recordset ‘Recordset
Dim ConnStr As String ‘Database connection string.
Dim Counter As Integer ‘Loop counter variable.

‘Create a connection string.

ConnStr = “Provider=sqloledb;” & _
“Data Source=WinServer;” & _
“Initial Catalog=Survey;” & _
“User Id=sa;Password=; “

‘Connect to the database.

Set Conn = New ADODB.Connection
Conn.Open ConnStr

‘Obtain the required recordset.

Set Recc = New ADODB.Recordset
Recc.Open “Food”, _
Conn, _
adOpenStatic, _
adLockOptimistic, _
adCmdTable

‘Create an array to hold the results. The database holds

‘8 Questions * 5 Answers Each, so we need 40 elements.
Dim MyData() As Integer
ReDim MyData(39)

‘Zero the array elements.

For Counter = 0 To 39
MyData(Counter) = 0
Next

‘Loop through the database and count the entries.

Do While Not Recc.EOF
MyData(Recc!LikeMeat) = MyData(Recc!LikeMeat) + 1
MyData(Recc!LikeVegetable + 5) = MyData(Recc!LikeVegetable + 5) + 1
MyData(Recc!LikeFruit + 10) = MyData(Recc!LikeFruit + 10) + 1
MyData(Recc!LikeGrain + 15) = MyData(Recc!LikeGrain + 15) + 1 Ch
MyData(Recc!DislikeMeat + 20) = MyData(Recc!DislikeMeat + 20) + 1
MyData(Recc!DislikeVegetable + 25) = _
MyData(Recc!DislikeVegetable + 25) + 1
7
MyData(Recc!DislikeFruit + 30) = MyData(Recc!DislikeFruit + 30) + 1
MyData(Recc!DislikeGrain + 35) = MyData(Recc!DislikeGrain + 35) + 1
Recc.MoveNext
218 Chapter 7 Creating Data Entry Forms and Surveys

Listing 7.3 Continued

Loop

‘Return the array of raw data.

GetFoodData = MyData

‘Close the recordset and database.

Recc.Close
Conn.Close

‘Clear the variables.

Set Recc = Nothing
Set Conn = Nothing
End Function

This component uses the same process for opening and closing both the connection and the
recordset, so we won’t discuss those issues again. Notice that this component outputs the
data to an array of integers. This technique wouldn’t work very well if the user wants to see
the comments as well. However, because this component is creating a summary of the data,
using an array works fine. (We’ll see later that there are caveats when using arrays; compati-
bility problems abound.)
The component creates a dynamic array, dimensions it for the task at hand, and zeroes every
entry in the array. Notice that there are 8 questions with 5 answers each. This means we
need 40 array elements (or 39 in a zero-based array). The component directly manipulates
the temporary array and adds the database entries using an index offset. It returns the raw
data to the client for further analysis.
You might wonder why the example doesn’t use a multidimensional array. Most SOAP
implementations don’t support multidimensional arrays, but many will support a single
dimension array. Using a single dimension does flatten the data and make it more difficult to
work with, but at least you can transfer it from client to server. None of the vendors that
produce SOAP toolkits intend to support multidimensional arrays as of this writing.
Theoretically, you can perform full data analysis anywhere in the application. Returning raw
data to the client, in this case, allows for local analysis. This means fewer trips to the server
over a slow connection. In addition, the customer will use local resources for the final analy-
sis, rather than rely on your server’s resources. The idea is to reduce bandwidth usage as
much as possible to ensure good application performance despite any latency that SOAP
might introduce.

Designing an Output Application

The survey analysis client will accept the raw data, perform any additional analysis, and pre-
sent it to the user. In this case, the survey client merely converts the raw input to percent-
ages and presents both to the user. Listing 7.4 shows the source code for the analysis client.
 Creating a Simple Survey Form 219

Listing 7.4 Analysis Client for Food Survey Source Code

Private Sub cmdGetSurvey_Click()
Dim MyArray() As Integer ‘Data array holding raw data.
Dim Counter As Integer ‘Loop counter variable.
Dim LikeMeatTotal As Integer ‘Total LikeMeat entries.
Dim LikeVegTotal As Integer ‘Total LikeVegetable entries.
Dim LikeFruitTotal As Integer ‘Total LikeFruit entries.
Dim LikeGrainTotal As Integer ‘Total LikeGrain entries.
Dim DislikeMeatTotal As Integer ‘Total DislikeMeat entries.
Dim DislikeVegTotal As Integer ‘Total DislikeVegetable entries.
Dim DislikeFruitTotal As Integer ‘Total DislikeFruit entries.
Dim DislikeGrainTotal As Integer ‘Total DislikeGrain entries.

‘Create the SOAP client.

Dim FoodSurvey As SoapClient
Dim ErrorMessage As String

‘Set up an error handler.

On Error GoTo ErrorHandler

‘Create the connection.

Set FoodSurvey = New SoapClient
FoodSurvey.mssoapinit _
“http://winserver/soapexamples/SurveyAnalysis/SurveyArray.WSDL”, _
“SurveyArray”, _
“FoodSoapPort”

‘Fill the data array with the current statistics.

MyArray = FoodSurvey.GetFoodData()

‘Release the object.

Set FoodSurvey = Nothing

‘Zero the total variables.

LikeMeatTotal = 0
LikeVegTotal = 0
LikeFruitTotal = 0
LikeGrainTotal = 0
DislikeMeatTotal = 0
DislikeVegTotal = 0
DislikeFruitTotal = 0
DislikeGrainTotal = 0

‘Display the raw data on screen.

For Counter = 0 To 4
LikeMeat(Counter).Caption = CStr(MyArray(Counter))
LikeMeatTotal = LikeMeatTotal + MyArray(Counter)
LikeVegetable(Counter).Caption = CStr(MyArray(Counter + 5))
LikeVegTotal = LikeVegTotal + MyArray(Counter)
LikeFruit(Counter).Caption = CStr(MyArray(Counter + 10))
Ch
LikeFruitTotal = LikeFruitTotal + MyArray(Counter)
LikeGrain(Counter).Caption = CStr(MyArray(Counter + 15))
LikeGrainTotal = LikeGrainTotal + MyArray(Counter)
7
220 Chapter 7 Creating Data Entry Forms and Surveys

Listing 7.4 Continued

DislikeMeat(Counter).Caption = CStr(MyArray(Counter + 20))
DislikeMeatTotal = DislikeMeatTotal + MyArray(Counter)
DislikeVegetable(Counter).Caption = CStr(MyArray(Counter + 25))
DislikeVegTotal = DislikeVegTotal + MyArray(Counter)
DislikeFruit(Counter).Caption = CStr(MyArray(Counter + 30))
DislikeFruitTotal = DislikeFruitTotal + MyArray(Counter)
DislikeGrain(Counter).Caption = CStr(MyArray(Counter + 35))
DislikeGrainTotal = DislikeGrainTotal + MyArray(Counter)
Next

‘Display the percentage data on screen.

For Counter = 0 To 4
LikeMeatPct(Counter).Caption = _
Left(CStr(MyArray(Counter) / LikeMeatTotal * 100), 4) & “%”
LikeVegetablePct(Counter).Caption = _
Left(CStr(MyArray(Counter + 5) / LikeVegTotal * 100), 4) & “%”
LikeFruitPct(Counter).Caption = _
Left(CStr(MyArray(Counter + 10) / LikeFruitTotal * 100), 4) & “%”
LikeGrainPct(Counter).Caption = _
Left(CStr(MyArray(Counter + 15) / LikeGrainTotal * 100), 4) & “%”
DislikeMeatPct(Counter).Caption = _
Left(CStr(MyArray(Counter + 20) / DislikeMeatTotal * 100), 4) & “%”
DislikeVegetablePct(Counter).Caption = _
Left(CStr(MyArray(Counter + 25) / DislikeVegTotal * 100), 4) & “%”
DislikeFruitPct(Counter).Caption = _
Left(CStr(MyArray(Counter + 30) / DislikeFruitTotal * 100), 4) & “%”
DislikeGrainPct(Counter).Caption = _
Left(CStr(MyArray(Counter + 35) / DislikeGrainTotal * 100), 4) & “%”
Next

‘We’re finished, so exit.

Exit Sub

‘Display a message when an error occurs.

ErrorHandler:

ErrorMessage = “Fault Code: “ + FoodSurvey.faultcode + vbCrLf + vbCrLf + _

“Fault String: “ + FoodSurvey.faultstring + vbCrLf + vbCrLf + _
“Fault Actor: “ + FoodSurvey.faultactor + vbCrLf + vbCrLf + _
“Detail: “ + FoodSurvey.detail
MsgBox ErrorMessage, vbExclamation
End Sub

Private Sub cmdQuit_Click()

‘Exit the application
End
End Sub

As you can see, the client creates a SOAP client, makes the method call, and releases the
SOAP client object. Most of the examples we’ve had so far have used the data returned from
the method call immediately. This application performs analysis on the data before present-
ing all of it, so it’s important to release the SOAP client immediately to free resources.
 Creating a Simple Data Entry Form Application 221

The client provides two display routines. The first unfolds the array and presents the raw
data onscreen. The second uses totals created during the raw data display to calculate data
percentages. It again unfolds the data, performs the calculation, and displays the result. A
multidimensional array would have made this process easier, but the flat presentation that
the single dimension array provides doesn’t add much processing to the code in this case.

Survey Application Data Processing Concerns

If you compare the size of this data transfer to a parse string, you’ll find that the string con-
sumes fewer resources, but requires more client-side code. The reason for the difference is
that SOAP sends the string as a single entity, but uses XML elements to transfer each array
element separately. Surveys tend to produce large amounts of data, so you might find that
you need to use the string presentation to obtain acceptable application performance.
This and other database clients might experience another problem when using the
Microsoft SOAP Toolkit (or other toolkits that rely on IIS). The example uses Internet
Server Application Programming Interface (ISAPI) instead of Active Server Page (ASP).
ISAPI has limits on the size message it can transfer. The Microsoft SOAP Toolkit automati-
cally sets this limit during installation to prevent denial of service (DOS) attacks by crackers.
You can change it using a registry setting,
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\MSSOAP\SOAPISAP. This key
has three values: MaxPostSize, NumThreads, and ObjCachedPerThread. Changing MaxPostSize
will allow you to send larger messages using ISAPI.

Testing the Output

One of the problems with using automatic code generators is that they don’t always produce the
best code. Our example uses an array that will normally work with any Microsoft SOAP Toolkit
client. However, you need to know that you might encounter compatibility problems when
using a WSDL file created with the WSDL Generator utility that ships with the Microsoft
SOAP Toolkit. Figure 7.14 shows the WSDL output of the SurveyAnalysis component.
Notice that the array dimensions are undefined. This will cause problems for some SOAP
implementations. You can hand edit the WSDL file to add dimensions as necessary, or you
can use a different WSDL generator. The lack of dimensions won’t cause a problem with
Microsoft clients.
The analysis client should run relatively quickly. It uses small data transfers to provide the
server with all of the survey data. Figure 7.15 shows the output from this application.

Creating a Simple Data Entry Form Application Ch

The data entry form application covers a lot of ground. For example, a survey is a type of
data entry form, albeit limited in scope. Some people also include full-fledged database
7
applications (discussed in Chapter 8) in this mix. However, there’s a middle ground applica-
tion that better defines the simple data entry form application. It’s the kind of application
where you expect two-way communication, but the communication is limited.
222 Chapter 7 Creating Data Entry Forms and Surveys

Figure 7.14
Look out for compati-
bility problems when
working with arrays.

Figure 7.15
The output from the
analysis client shows
all of the statistics for
the food survey.
 Creating a Simple Data Entry Form Application 223

In this section, we’ll look at a data entry form where the user enters information and
receives a response based on his input. The user enters data, but can’t modify the input in
any way. The conversation is two-way because the server provides a response based on the
user input. The following sections will show you how to put this example together.

Creating the Database

This example uses two databases. The first appears in Figure 7.16. It contains the user’s
input. Notice that the table uses a primary key of the user’s first and last name to ensure
there aren’t any duplicates. You’d need to make the primary key more distinct in a real-
world application because many people have the same first and last name.

Figure 7.16
The user input is lim-
ited to name, age, sex,
a like, and a dislike.

The second table appears in Figure 7.17. It contains two fields. The first is a number used to
access a prediction, and the second is the prediction string. The user won’t have access to
this table—the administrator will control the prediction strings.

Figure 7.17
This prediction
string table is under
the administrator’s
control.

Designing the Server-Side Component

The server-side component will accept input from the user, then use that input to construct
an output string. The output string will also contain data taken from another table. Users
commonly rely on data entry applications to gain access to information they want for a par-
ticular purpose. Filling out a survey or user identification form might garner access to a pri-
vate Web site. Listing 7.5 shows the source code for this example.

Listing 7.5 Data Entry Component Source Code

Public Function GetPrediction(ByVal First As String, _ Ch
ByVal Last As String, _
ByVal Age As Integer, _ 7
ByVal Sex As String, _
ByVal Likes As Integer, _
ByVal Dislikes As String) As String
224 Chapter 7 Creating Data Entry Forms and Surveys

Listing 7.5 Continued

Dim Conn As ADODB.Connection ‘Database connection.
Dim Recc As ADODB.Recordset ‘Recordset
Dim ConnStr As String ‘Database connection string.
Dim Output As String ‘Prediction string.
Dim RecCount As Integer ‘Total number of prediction records.

‘Create a connection string.

ConnStr = “Provider=sqloledb;” & _
“Data Source=WinServer;” & _
“Initial Catalog=DataEntry;” & _
“User Id=sa;Password=; “

‘Connect to the database.

Set Conn = New ADODB.Connection
Conn.Open ConnStr

‘Obtain the user input recordset.

Set Recc = New ADODB.Recordset
Recc.Open “UserInput”, _
Conn, _
adOpenStatic, _
adLockOptimistic, _
adCmdTable

‘Enter the data.

Recc.AddNew
Recc!First_Name = First
Recc!Last_Name = Last
Recc!Age = Age
Recc!Sex = Sex
Recc!Likes = Likes
Recc!Dislikes = Dislikes
Recc.Update

‘Create the output string.

Output = “Hello “ & First & “ “ & Last
If Age > 0 And Age < 40 Then
Output = Output & “. You’re a young “
ElseIf Age >= 40 And Age < 65 Then
Output = Output & “. You’re a middle aged “
ElseIf Age >= 65 Then
Output = Output & “. You’re an old “
Else
Output = Output & “. You’re a mysterious “
End If

If UCase(Sex) = “M” Then

Output = Output & “man.”
ElseIf UCase(Sex) = “F” Then
Output = Output & “woman.”
Else
Output = Output & “person of unknown gender.”
End If

‘Close the user input recordset.

Recc.Close
 Creating a Simple Data Entry Form Application 225

Listing 7.5 Continued

Set Recc = Nothing

‘Open the prediction recordet.

Set Recc = New ADODB.Recordset
Recc.Open “Prediction”, _
Conn, _
adOpenStatic, _
adLockOptimistic, _
adCmdTable

‘Determine the number of records.

Recc.MoveLast
RecCount = Recc!PNumber

‘Find a random location in the prediction database.

Randomize
RecCount = Int((RecCount * Rnd) + 1)
Recc.MoveFirst
Recc.Find (“PNumber = “ & RecCount)

‘Finish creating the output string.

Output = Output & vbCrLf & “Your prediction is: “ & Recc!PString
GetPrediction = Output

‘Close the recordset and database.

Recc.Close
Conn.Close

‘Clear the variables.

Set Recc = Nothing
Set Conn = Nothing

End Function

As you can see, this component begins by creating a connection to the database. We require
access to two tables, but not at the same time. This allows use of the same variable for both
accesses, as shown in the source code. Notice that in the first case, the application adds a
record to the database without performing further analysis. In the second case, the compo-
nent queries a random value from the database without adding anything new.
The remainder of the code in this component performs analysis on the input. Data entry
applications often perform this type of service (of course the analysis is a little more complex
than the code shown here). In this case, the user obtains feedback on the input.

Creating a Data Entry Client

The data entry client application is responsible for gathering data from the user, ensuring Ch
the data is of the right type, and sending it along to the server. It also presents the output 7
to the user. In this case, the output is a simple message box. Listing 7.6 shows the code for
this example.
226 Chapter 7 Creating Data Entry Forms and Surveys

Listing 7.6 Data Entry Client Source Code

Private Sub cmdPredict_Click()
Dim SendAge As Integer

‘Create the SOAP client.

Dim MyPrediction As SoapClient
Dim ErrorMessage As String

‘Set up an error handler.

On Error GoTo ErrorHandler

‘Create the connection.

Set MyPrediction = New SoapClient
MyPrediction.mssoapinit _
“http://winserver/soapexamples/DataEntry/DataEntry.WSDL”, _
“DataEntry”, _
“PredictionSoapPort”

If Len(txtAge.Text) = 0 Or Not IsNumeric(txtAge.Text) Then

SendAge = 0
Else
SendAge = txtAge.Text
End If

MsgBox MyPrediction.GetPrediction(txtFirst.Text, _
txtLast.Text, _
SendAge, _
comboSex.Text, _
comboLikes.ListIndex, _
comboDislikes.ListIndex)

‘We’re finished, so exit.

Exit Sub

‘Display a message when an error occurs.

ErrorHandler:

ErrorMessage = “Fault Code: “ + MyPrediction.faultcode + vbCrLf + vbCrLf + _

“Fault String: “ + MyPrediction.faultstring + vbCrLf + vbCrLf + _
“Fault Actor: “ + MyPrediction.faultactor + vbCrLf + vbCrLf + _
“Detail: “ + MyPrediction.detail
MsgBox ErrorMessage, vbExclamation
End Sub

Private Sub cmdQuit_Click()

‘Exit the application.
End
End Sub

Private Sub Form_Load()

comboLikes.ListIndex = 0
comboDislikes.ListIndex = 0
End Sub
 Security, Privacy, Performance, and Reliability Issues 227

Notice that we don’t provide default values for either age or sex. In addition, you’ll notice
that the component code in Listing 7.5 handles these two fields with some level of discretion.
Finally, the database design allows for NULL values in these two fields. All of these coding
practices reflect the need to protect the user’s privacy. It’s a consideration for all SOAP appli-
cations because they provide some type of online service that might appear in the public eye.
The code does check for problems in the age field. It’ll provide a value if the user doesn’t
enter an age—one that the application will associate with a NULL value. Likewise, the appli-
cation assumes the user didn’t want to add a value if he supplies a string rather than a num-
ber in the age field.
None of the other fields requires a validity check. The application must assume that the user
knows his name. The sex, likes, and dislikes fields all use combo boxes, which allow us to keep
values in check without adding code. Notice that the code automatically sets the likes and dis-
likes values to default values, but leaves the sex field alone so isn’t any assumption of gender.

Testing the Data Entry Application

It’s time to test the application. You’ll need to register the component on the server and
create the usual WSDL file. Figure 7.18 shows some sample output. Notice that the string
contains a mix of user input and data from the prediction database.

Figure 7.18
The output of the
Data Entry application
is a restatement of
the input data and
a prediction.

It’s important to include a range of input values during the testing process. Try arcane com-
binations of input. For example, this application doesn’t test for blank fields, something that
could cause problems. It should also check for duplicate entries in the database, but SQL
Server does take care of this problem.

Security, Privacy, Performance, and

Reliability Issues
It might seem a bit strange to place the four topics in the heading for this section together,
but you really do need to consider them as a group. Changing one item normally affects the
other three in a negative way. For example, when you increase privacy, you decrease what
you know about those visiting your Web site. As a result, your ability to search for potential Ch
security flaws decreases. Enhancing privacy also means using more system resources, so per- 7
formance takes a hit. Finally, adding more processing to your system increases the number
of failure points, which decreases reliability. In short, you can’t make one change without
changing your entire setup.
228 Chapter 7 Creating Data Entry Forms and Surveys

Interactions are nothing new for developers. Most of us have had to deal with code interac-
tions as part of every programming exercise. The code you create in one area invariably pro-
duces an effect somewhere else—sometimes in a way that you never anticipated. However,
SOAP represents a new set of interactions to consider because the conditions are different.
SOAP operates with the Internet, making it possible that outside sources will affect your
system. In addition, SOAP features new functionality that you haven’t had in the past.
The following sections will look at security, privacy, performance, and reliability from a
SOAP perspective. We’ll consider how this new environment will affect how you design and
write an application. These sections will also discuss a difference in ideology—the way you
approach a development problem.

Security
One of the most important features for any application today is security. Read some of the
trade press magazines and you’ll discover the latest cracker attack on a large company. Some
of these attacks are so severe that they attract attention in large national magazines and
newspapers. In short, the problem is real and it’s significant. As a developer, you’re right in
the middle of the current problems with security.
In days past, a developer could rely on the operating system and third-party products to take
care of security problems. Local server settings are often enough to ensure safe computing
in a LAN environment. These sources still help, but the developer often has to do more to
ensure the application remains safe. The question, of course, is how safe is safe enough?
Every time you add a new security feature, you further obscure the application from public
view. However, security features consume resources that affect the performance of your
application. In addition, each new security feature is a failure point, so adding a security fea-
ture reduces the reliability of your application. In short, you have to reach a balance
between enough security and adequate performance and reliability.
Adding security features can affect privacy. As you add features, the user normally has to
provide more information for identification purposes. After all, one of the main goals of
security is to ensure that you know who is at the other end of the wire. This is the negative
impact of privacy. An anonymous user can’t access your site, so on some level he has to give
up some privacy to use a resource on your server.
Fortunately, security also has a positive impact on privacy. For one thing, the user worries
less about someone else gaining access to his information. By reducing the chances of a
cracker breaking into your server, you also increase the possibility that no one will find out
about your user base.
Of course, everyone should know that there aren’t any completely secure systems. If a
cracker has enough time, he will eventually break into your system. Security increases the
time required for a break-in and allows you to detect a break-in with greater ease, but it
won’t prevent a break-in from happening. As part of your security strategy, you need to con-
sider how much management the system will require and use that as the determining factor
for the security, performance, and reliability trade-offs.
 Security, Privacy, Performance, and Reliability Issues 229

From a SOAP perspective, adding security to an application is difficult. SOAP provides no

built-in security, and using external security is a halfway measure at best. You also have to
consider the severe impact that adding layers of security will have on your application.
SOAP already bloats the size of data by adding XML tags to it. These tags are necessary in
the absence of the information that binary data transfer provides. Encrypting the data
increases its size even more. In short, you need to use larger numbers when predicting the
impact of security on the performance and reliability of your application.
How large of an impact could using security and SOAP together have on your data? Some
people report that their data increases by as much as 600% when converting it to SOAP, and
using SSL increases the size another 150%. This is probably a worst-case scenario, and few
real-world statistics exist at this moment. However, given these numbers, a 1,000-byte data
transfer suddenly consumes 9,000-bytes when transferred using both SOAP and SSL.
(These figures are based on several sources, including vendors and newsgroup messages—
there aren’t any formal tests available at the time of this writing.)

Privacy
Every country in the world is becoming serious about the privacy of individuals. The reason
is simple. Not only does protecting individual security make good political sense, but people
do have a right to privacy. From a corporate goodwill perspective, it’s essential to protect
customer data. You won’t maintain a customer base for long if your company gains a reputa-
tion for giving out customer information at the drop of a hat, especially in countries where
privacy policies tend toward strict confidentiality.
Every time you collect survey information from an unknown party, you risk divulging infor-
mation about that individual that could have unforeseen repercussions. Given that privacy is
such an important issue, you don’t want to risk giving out information, even if it’s the result
of a cracker attack on your system. In short, the best way to conduct a survey is to gather
the information anonymously. However, collecting information in this way almost guaran-
tees that someone will try to contaminate the survey by stuffing the ballot box (taking the
survey more than one time).

One favorite method of overcoming the problem of identification is to ask for the user’s
e-mail address and allow just one entry per e-mail address. This is the least invasive way
to assure the survey is unique. If a cracker does break into your system, he’ll receive an
e-mail address that’s somewhat public and not any personal information about the
respondent. The e-mail address also allows you to perform spot checks and ensure that
the person actually participated in the survey. You don’t have to know a name, address,
or other information to use this technique, so it’s safe from a privacy perspective.
Ch

7
230 Chapter 7 Creating Data Entry Forms and Surveys

Privacy issues are difficult on other fronts as well. The less information you collect from
someone before allowing access to your system, the higher the chance that he’ll use the
opportunity to circumvent your security. Increasing privacy tends to affect security in a neg-
ative manner because security principles dictate knowing as much as possible about the per-
son you’re dealing with over any connection.
Adding privacy features to your application also increases processing load. Any additional
processing uses resources that you could use for other purposes. Consequently, increasing
privacy normally incurs a performance loss. Fortunately, this isn’t always the case.
The amount of performance loss depends on how you implement privacy. Unlike security,
which you want to automate component, you can implement privacy through policies and
other human activity changes. For example, you could create a company policy that requires
personal telephone contact for anything other than an anonymous survey. Your application
could actually see a performance increase if company policy dictates purging all personal
information at regular intervals. Removing individual surveys and preserving just the raw
tallies after a longer storage period could allow further performance increase because the
databases holding the data contain fewer records.
Privacy usually affects reliability in a negative way, even if you’re implementing privacy
using policies. No matter which way you look at it, adding privacy features to an application
means adding more failure points. In fact, using policies actually decreases reliability even
more than adding software components because policies require interpretation on the part
of the implementer.

Performance
Many developers are finding that the myth of unlimited bandwidth and resources is just
that—a myth. As soon as a company adds more bandwidth to a network, management and
users find new ways to use that bandwidth. A network that feels comfortable for a while
after an upgrade reverts to providing slow access in a short time because of increased band-
width usage. Likewise, a larger hard drive invites users to store more data, rather than
archive it. Any increase in storage will almost certainly disappear. In short, performance is
always going to be a problem that developers need to solve.
One of the dichotomies that developers must consider is that society will continue to demand
new application features while users complain about the problems of code bloat. For example,
we’ve discussed increased security and privacy features in this chapter. Users don’t dictate these
requirements; corporate management and government demand that developers add them. The
developer is caught in the middle trying to decide how best to implement required features in
such a way that performance isn’t affected, but application requirements are satisfied.
SOAP applications suffer from code bloat more than desktop applications because they require
extra security and privacy. An open application requires more code to ensure a safe user session.
Unfortunately, SOAP applications also have fewer resources because Internet connections aren’t
as fast as the LAN in most cases. As you can see, if you spend a little time tuning performance
on a desktop application, you’ll spend a lot of time tuning SOAP application performance.
 Security, Privacy, Performance, and Reliability Issues 231

(There are exceptions to every rule—see the “When SOAP Is Faster” section of Chapter 1, “An
Overview of SOAP,” for some ideas on when SOAP actually provides better performance than a
desktop application.)
The best way to view this situation is a problem of trade-offs. If you want maximum perfor-
mance from your application, you have to be willing to design your application carefully.
SOAP can help in this regard if your remember some simple SOAP rules. For example,
SOAP is a one-way transfer protocol. Sure, you can use a request/reply format to your mes-
sages, but there isn’t a prolonged session to consider. This means that you can add security
to just the data transfers that contain sensitive information.
Likewise, establish user identity after using security e-mail or some other means. Provide
the user with a keyword or other unique means of identification he can use. You can ensure
privacy from prying cracker eyes by using code and keeping user records on a server that
isn’t directly accessible from the Internet connection. Using this technique means the
impact on performance will remain small while keeping user privacy high.

Many developers are used to using long function and variable names. Use short names
whenever possible for public functions and variables. This decreases the size of the
SOAP tags and enhances application performance.

Reliability
Application reliability is an ever-increasing concern as companies begin competing in a
world market. The term, “24/7” is as much a part of our vocabulary now as Internet. Most
corporations interested in SOAP application development have a requirement for applica-
tions that run 24 hours a day, 7 days a week without failure. Hardware engineers have gone
to lengths to ensure that a hardware failure means a loss of performance, not down time.
Many vendors now strive for five 9s (99.999%) of up time.
As mentioned throughout this section, any time you add a new module to your application,
you reduce reliability. That’s because reliability prediction bases the reliability of your sys-
tem on the number of failure points (application modules) as well as the reliability of a par-
ticular module.
However, many of the modules in your application aren’t under your direct control. For
example, you’ll likely use another vendor’s XML parser instead of one created by your in-
house staff. Likewise, we’ve used SOAP toolkits created by other companies throughout the
book. It makes sense from a developer productivity perspective to use third-party products.
In addition to third-party application modules, you also have the activities of others to con- Ch
sider. The Internet connection you use is under the ISP’s control. Crackers can probe mes- 7
sages as they move between locations on the Internet. Even nature gets into the act by
adding noise to the line and interrupting the signal.
232 Chapter 7 Creating Data Entry Forms and Surveys

Obviously, you have no control over conditions outside the application or modules you
didn’t design, but you can control application elements you did design. The following list
provides some ideas for increasing the reliability of SOAP applications:
■ Keep the number of modules in your application small. Use modules as needed to provide
performance, privacy, and security benefits, but try combining modules whenever possible.
■ Investigate compatibility issues between platforms before you begin the design process,
then compensate within your application code.
■ Test modules individually and integrated into the application. Modules that you test
fully will cause fewer reliability problems.
■ Look for places where you can reduce the number of processing steps. This increases
both performance and reliability. Of course, you need to consider the nature of SOAP
applications and add both security and privacy features as needed.
■ Use standards-based programming whenever possible. This ensures the remote site will
understand the messages you send. This is especially important for security and will
probably become important for privacy issues as well.

Handling Data Entry and Survey Form Errors

Data entry and survey form applications have one thing in common: the end user is unlikely
to know as much as the average desktop user about error messages. In addition, these users
aren’t in a position to do much about the error even if they do know enough to understand
what it means. In short, you need a new way to handle application errors.
Providing a means to record errors on the server for the component isn’t hard. All you need
to do is trap the error and record the event in a log. When working with a survey applica-
tion, you use the event log in place of a message to the user. Try to record the incoming data
as part of the event message so that you preserve the survey if possible. This allows someone
to enter the survey again later after the component problem is resolved.
In previous chapters, we used the SOAP Fault message to handle error messages. However,
consider the user’s reaction if he sees a message that states the server isn’t responding or has
experienced some type of processing difficulty. The user might be at a kiosk and won’t care
about the server. All the user knows is that the application isn’t providing the required infor-
mation or allowing him to complete an action. Considering the fact that you’re asking the
user for a service (answering survey questions, for example), he might not care and give up
after the first problem.
The inability to send a message back to the user of a survey application or the need to send
something comprehensible to the user of a data entry application presents challenges. If
you’re using the Microsoft SOAP Toolkit, it means that you have to use the ASP method of
working with the SOAP message because the ISAPI method lacks the means to capture the
fault message. Edit the ASP file by hand to create an event log entry for the administrator,
rather than send the fault message back to the client. When working with a data entry
 Using Templates for Quick Forms 233

application, you’ll need to come up with a simple message that everyone will understand,
but that doesn’t necessarily provide complete error information.
We’ve covered the Web server and the component. What happens if the error occurs
between the client and the Web server? For example, the application might receive a time-
out message if the server takes too long to respond. Depending on the SOAP toolkit you
use, it might be possible to add scripts to handle some of the common problems such as
time-outs. Unfortunately, recording an event log entry is going to be difficult or impossible.
In this case, you might want to provide a telephone number as part of the user message.
This will allow the user to call someone to notify them of the problem without burdening
the user with information he doesn’t want to know.

Using Templates for Quick Forms

Ultimately, the need for a user interface ties SOAP to desktop or browser applications
somewhere along the way. The need to present the user with an interface means writing
code that displays labels and other visual cues that humans need to make sense of computer
data. You might find that you’re producing forms that vary only slightly because of the kind
of data you’re sharing or the way that your application works.
Surveys and data entry applications are well suited to use templates. A template describes
those common elements and allows the application to describe the rest. Consider templates
as the latest form of “fill-in-the-blank” technology.
Templates aren’t new. Web servers use templates regularly. One of the best reasons to use
frames is to allow a developer to place common elements in set locations and other windows
for changing data. Many Web servers even support the concept of headers and footers to
display common data without resorting to frames. The addition of technologies such as cas-
cading stylesheets (CSS) is the result of developer reuse requirements.
SOAP differs from many Web applications in that it needs these common elements on the
user’s machine because it uses client-side processing. In other words, creating an ASP that
creates a Web page based on user input probably won’t do the job unless you’re using the
ASP to create the initial page. The application should also include client-side features that
rely on templates so the application transfers the least amount of information to and from
the server. Remember that a SOAP application is client-based, not server-based. The user
isn’t going to download the SOAP application every time as he would a Web page.
Desktop applications present a different challenge. Most desktop applications rely on prede-
fined forms and never use any form of template. You can overcome this problem in a num-
ber of ways. For example, programming language products, such as Visual C++, allow you to
combine HTML with standard desktop applications. In short, you could use HTML tem- Ch
plates in the same way that you would with a Web-based application. 7
The bottom line is that templates can save a programmer time and create a more consistent
look for your application. Implementing template technology on the client isn’t difficult, but
it will require you to rethink some of the ways you’ve worked with templates in the past.
234 Chapter 7 Creating Data Entry Forms and Surveys

Using MIME for SOAP Applications

As SOAP applications become more complex and data handling needs increase, we’ll eventually
need ways to transfer data that doesn’t meet the XML specification requirements. The real
world contains data types that cause developers to pull their hair out by the handful. Transferring
data from older databases on mainframes could prove especially difficult. In short, the data types
that the XML specification defines are fine as a start, but hardly the end of the discussion.
One of the big problems that SOAP vendors face right now is the issue of working with
unwieldy data types. SOAP is essentially text, which limits the number of ways to translate
data. The answer that everyone is using right now is Base64 encoding. Of course, trying to
implement this as part of a SOAP message is difficult for even the best programmer. The
Multipurpose Internet Mail Extensions (MIME) attachments described in the
“Understanding SOAP Attachments” section of Chapter 2, “SOAP in Theory,” promise an
easier way to handle these data types. Using MIME means that you can separate the SOAP
message from the peculiar data you ask it to deliver to the client. Images no longer pose a
problem because they appear in a separate part of the message.

You currently have to use attachments or Base64 encoding to transfer images when
working with SOAP. The reason is simple: XML doesn’t support an image data type, and
SOAP relies on XML standards support. A future version of SOAP might include a native
image type. The Association for Information and Image Management International
(AIIM) (http://www.aiim.org) is working on a new image standard for XML. The
International Organization for Standardization (ISO)(http://www.iso.ch/) is sup-
porting the effort as part of TC171/SC2/WG2. AIIM also has plans to work closely with
the World Wide Web Consortium (W3C) (http://www.w3.org/). It’s helpful to keep
track of standards such as this one because they greatly reduce the effort required to
create database applications.

My original plan for this chapter included showing how to use MIME to add pizzazz to a
data entry application. Imagine being able to send photographs or other hard-to-handle data
across the Internet using SOAP. Unfortunately, no one has a SOAP toolkit that will create a
MIME message at the time of this writing. The specification defines the techniques for
working with MIME, but the required support is sadly lacking. One vendor went so far as to
say that MIME support would have to wait until enough user comments showed a need.
I did learn a few things about using MIME with SOAP during the course of planning this
chapter. For example, some developers have decided not to wait and are writing their own
SOAP implementations that support MIME. Working with these talented individuals has
allowed me to draw the conclusions outlined in the sections that follow.

Special Considerations for MIME Messages

Implementing MIME for SOAP will require two levels of changes. The first is to the SOAP
toolkit that you use to create applications. SOAP toolkit vendors will find it challenging to
add MIME support because MIME relies on helper applications to do its job. When you
 Using MIME for SOAP Applications 235

work with an e-mail or browser application, you can define these helpers. Using MIME
seems automatic because the application list is readily available. However, SOAP doesn’t
provide a user interface as such, so where do you configure the helper application list?
Using the client’s list of helpers will work on the client machine, but not on the server.
Another problem that SOAP toolkit vendors will have to solve is the issue of attachment
compatibility. For example, it’s relatively easy to say that users will require some type of
image, sound, and multimedia support, but vendors will have to figure out which file types
SOAP should support. The image files on Macintosh are different than those on a PC (and
they’re stored differently as well).
MIME support will also require extensions to the XML parser on most machines. The ven-
dors will have to teach the parser the difference between the SOAP and the attachment por-
tion of the message. The parser might also face difficulties in parsing SOAP messages that
refer to attachments so the user can interact with them. The technology is already available
in part, but the implementation lags behind progress in other areas of SOAP.
The second level of change is developer use of SOAP toolkits. It’s unlikely that any MIME
implementation will rely on the high-level API that most developers prefer. This means that
developers will spend more time building messages by hand. Automation has certain advan-
tages. It’s important to plan for delivery delays due to the increased time required to use the
low-level API.
Developers will also spend more time considering cross-platform compatibility issues. Even
if the SOAP toolkit vendor provides a great parser and clearly defined policies for attach-
ment use, it’s up to the developer to resolve the differences between the SOAP toolkit capa-
bility and real-world requirements. For example, the developer will need to provide a means
for placing an image found in an attachment within a SQL database.

Where MIME Support Is Going

You can’t use MIME today, and given the number of issues vendors must resolve before
developers can start using MIME, you probably won’t see MIME support for a while. Some
vendors see MIME as an unnecessary add-on, while others want to work on other SOAP
requirements first. Eventually, you’ll be able to use MIME to transfer some types of data
using SOAP. The technology already appears in other parts of the Internet and it offers too
many advantages to ignore.

The W3C has created a specification for working with SOAP attachments (http://www.
w3.org/TR/SOAP-attachments). This information is in addition to the material that
we discussed in the “Understanding SOAP Attachments” section of Chapter 2. Despite the
number of documents that discuss SOAP attachments, no one has implemented them Ch
fully yet. Several vendors are working on including SOAP attachments at the time of writ-
ing, so you might find a toolkit that includes the required support.
7

The short answer is that MIME will have to wait until other parts of SOAP are better
defined, vendors have finally solved some of those compatibility problems, and pressing
236 Chapter 7 Creating Data Entry Forms and Surveys

issues such as security are resolved. SOAP is a new technology and it will require time to
define fully.
MIME will also have to evolve. The current implementation of MIME defines multi-part
message delivery for applications such as e-mail and browsers. Vendors rely on the user to
fill in many of the gaps required to make MIME work. They won’t have that luxury when
working with SOAP. It may be that we’ll eventually end up with something like MIME, only
better defined and explicit for use with SOAP. Only time will tell.

Project
The best way to learn how to create surveys and data entry forms is to create a few projects
of your own. Creating a survey is much easier than a data entry form when working with
SOAP because you only need to consider one-way data transfers. Surveys come in a wide
range of forms, providing you with a chance to experiment at several levels. Use the follow-
ing steps to create several survey SOAP applications of your own.
1. Design a database. Make sure you consider privacy and security issues as part of the
design. In many cases, making the survey anonymous works just fine. Try using an e-mail
address in situations where you need to identify the user in the least invasive way possible.
2. Create a server-side survey component. Make sure this component provides only one-
way access. It should only allow the user to pass survey information to the server; the
server shouldn’t send any information back. Adding a success message to the client
application tends to reduce the need for server communication. It’s also important to
create a special account for survey access. Use a different account for each survey to
ensure maximum protection.
3. Create the survey form. It often helps to build a local client first, then use the code you
create as the basis for a remote client. Ensure you provide full error message support
because the user will likely have little computer experience.
4. Add a minimum of nine records to your database using the survey application. This
allows you to test the application and build records for analysis. Make sure your data-
base contains enough entry types that you can test application functionality fully.
5. Create a server-side analysis component. Make sure this component provides read-only
access to the database. This ensures that data integrity remains high and reduces the
chance that the client will accidentally corrupt the database. Create a special account
with read-only access for this survey analysis. In some cases, you might have to provide
supervisor access using a separate component and account for managers at the client.
Adding filtering and indexing capabilities for the analysis component is one way to
improve performance because the client will only download needed data.
6. Design a survey analysis form. In fact, you might need to design several forms to meet
client needs. In addition to raw data, provide outputs showing interpreted data. For
example, you might want to provide percentage for each raw data output. Many analysis
programs also provide graphics, such as pie charts, to ensure the client receives maxi-
mum benefit from the data.
 CHAPTER
8
Providing Remote Database
Access
In this chapter
Remote Database Application Uses and Concerns 239
Developer Shortcuts for Remote Database Applications 243
Using Complex Data Types 245
Defining the SQL Server Database 255
Creating the Server-Side Component 255
Creating a Middle-Tier Component 263
Creating the Client-Side Application 267
Testing the Application 272
Quick Fixes for Remote Database Applications 274
Addressing Transaction Issues 278
Troubleshooting 279
238 Chapter 8 Providing Remote Database Access

Databases form the core of most, if not all, businesses. Everyone needs to store data for
future use. The data a business owns defines it and enables it to compete. Businesses also
sell data and share it with partners. Customers rely on the data a business owns in the form
of catalogs to make purchases. In short, it’s natural that you’d want to provide some form for
remote database access using SOAP. Using SOAP will allow a business to exchange data
with more businesses than ever before in more ways than anyone would have thought possi-
ble even a few years ago.
This chapter is going to help you understand the SOAP-to-database connection. You’ll find
that SOAP brings with it special requirements for database access. Directly accessing data in
a database with SOAP is unwise, and we’ll explore the reasons why in the first section of this
chapter, “Remote Database Application Uses and Concerns.” This section will also explore
some of the better uses for databases with SOAP and some of the uses you should avoid.
The second section of this chapter, “Developer Shortcuts for Remote Database
Applications” will help you get your applications finished faster. We’ll look at techniques
you can use to avoid development pitfalls when developing SOAP applications. For exam-
ple, SOAP normally requires additional “buffer” components that you might not use in
most situations.
The third section of this chapter, “Using Complex Data Types,” contains a special example
showing how to use complex data types. The Microsoft SOAP Toolkit doesn’t provide the
best support for this type of example, so we’ll use pocketSOAP instead. This example does
point out the need for innovation when working with SOAP in its current state because not
every toolkit supports every requirement. You’ll find that compatibility requirements
become even more critical as the complexity of the application increases.
The fourth (“Defining the SQL Server Database”), fifth (“Creating the Server-Side
Component”), sixth (“Creating a Middle-Tier Component”), and seventh (“Creating the
Client-Side Application”) sections show you how to create a SOAP database application.
You’ll immediately see that some development tasks are the same as any other environment.
For example, the construction of a database doesn’t change simply because you’re using
SOAP. In other cases, such as the middle-tier component, you’ll see how to work within the
limits of SOAP.
After you have the application put together, it’s time to test it. The eighth section of this
chapter, “Testing the Application,” will tell you how to test SOAP applications effectively.
It’s important to remember that you have less control over a SOAP application in some ways
because the client and server aren’t directly connected. You also need to test more failure
conditions, such as the loss of a connection.
The ninth section of this chapter, “Quick Fixes for Remote Database Applications,” tells you
how to repair your database application when it fails. Many developers are under the
impression that software doesn’t fail. In the strictest sense of the word, they’re correct
because software doesn’t wear out. However, applications can fail in many other ways, and
this section tells you about the problems that you’re most likely to see. We’ll look at data-
base applications in general and SOAP applications in particular.
 Remote Database Application Uses and Concerns 239

I mentioned earlier that data defines your business. This means that data loss can cripple
your business in ways that you might not know about until the loss occurs. One solution to
the problem is to ensure that data loss doesn’t occur, or at least becomes a rare event. Most
database applications today rely on transactions to ensure that data transfers occur safely.
Ch
Unfortunately, SOAP doesn’t provide a native means for handling transactions. We’ll talk
about this issue in the “Addressing Transaction Issues” section of this chapter. 8
The final section of this chapter, “Troubleshooting,” tells you how to troubleshoot SOAP
database applications based on everything you’ve learned in the chapter. We’ll go through
some problem scenarios that you’re likely to see so that you learn to recognize the source of
failures immediately.

Remote Database Application Uses and Concerns

Database management forms the core of most businesses. Each business-specific activity
revolves around some form of data storage and retrieval. Some developers work on database
management systems exclusively because the demand for these applications is so great.
Databases come in every size and complexity. It’s little wonder then that the ability of SOAP
to handle database management tasks is high on the list of developer concerns.

Database applications produce copious amounts of data. SOAP (actually XML in general)
increases the size of that data by wrapping it in tags. It doesn’t take much to figure out that
a SOAP application could overwhelm your server’s storage capacity in short order unless
you’re very careful about the methods used to store the information. One way to reduce
the effect of this problem is to compress the data. Unfortunately, standard compression
techniques leave the document object model (DOM) unreadable, which makes the data
inaccessible. In short, standard compression reduces the functionality of SOAP. Fortunately,
there’s a solution to the problem in the form of XMLZip (http://www.xmls.com/
resources/xmlzip.xml?id=resources_xmlzip). You can get XMLZip for the price
of a download. This product compresses your SOAP data and still allows applications to
access it as normal. You can get XMLZip for the Windows NT/2000, Solaris, and Linux plat-
forms. We’ll discuss this product more in Appendix C, “Third-Party Tool Reference.”

This section of the chapter isn’t going to tell you that SOAP is the greatest protocol in the
world for handling database application needs. In fact, SOAP is only useful when creating
certain classes of database applications. The first section that follows will tell you why SOAP
is less than perfect for database application needs. The second section will tell you where
SOAP excels, and which applications can use SOAP successfully. The bottom line is that
there are some situations where online processing of your data won’t work because the tech-
nology required to implement a solution doesn’t exist. SOAP is simply a first step into the
larger application development world.
240 Chapter 8 Providing Remote Database Access

Concerns
Let me say up front that the use of SOAP for any kind of database application concerns
many developers. The problem, according to purists, is that SOAP brings with it too many
limitations to provide effective database support. About half of the people I talked with
agreed that five issues hinder the use of SOAP for database applications.
■ Security: SOAP doesn’t provide native security. Adding security separately isn’t good
enough, according to some experts, because the use of two protocol levels leaves the
data exposed for some period of time (albeit a small one). In most cases, the security
provided by protocols such as Secure Sockets Layer (SSL) will work fine for non-critical,
but sensitive, data transfers.
■ Transactions: You won’t find transactional support in SOAP, and many developers feel that
it’s going to be impossible to add it. (An effort is being made now to add transactions, but
as of this writing, the committee studying the problems hasn’t made much progress.) This
is a significant problem, and one that should probably keep you from using SOAP for
transferring critical data. However, not all data transmissions on a LAN require transac-
tions, and it’s certain that not all data transmissions on the Internet will require them
either. The point is to choose the kinds of SOAP applications you create carefully.
■ Connectivity: Unlike your LAN, a SOAP connection can become broken at any time.
Because SOAP uses Hypertext Transfer Protocol (HTTP) or Simple Mail Transfer
Protocol (SMTP) as a transport protocol, you’ll find that the server doesn’t notice the con-
nection break immediately and that data loss can occur. SOAP doesn’t provide the robust
error handling of protocols such as Distributed Component Object Model (DCOM) and
Common Object Request Broker Architecture (CORBA). The use of a buffer component
such as the one shown in this chapter can help overcome this particular problem.
■ Data Type Translation: As of this writing, the SOAP specification contains little in the
way of support for complex data types. The specification also lacks information on
translating complex types to the simple types that the specification does support. Tool
vendors are currently trying to address this issue, but it seems that each tool vendor has
a different vision for data translation. Ensure that the SOAP toolkits you use for your
applications support the same data types and data translation schemes, and you can
probably avoid this problem.
■ Scalability: Database applications need to support thousands of users in many cases.
SOAP detractors assume that all of these users will rely on a SOAP connection to work
with data, and that the limitations of this protocol will keep the application from scaling
well. The verbose nature of the data exchange is also a source of concern because many
servers have a hard time keeping up with binary protocols they don’t need to translate.
The reality is that scalability could become a problem if you attempt to use SOAP in
more ways than the protocol will actually support. The best approach is to use SOAP
for remote communication as needed and to continue to rely on binary protocols for
local and secure remote communication.

These five issues would keep some developers from using SOAP for database applications at
all. You have to wonder why they’d even invest the time learning SOAP because it’s no
 Remote Database Application Uses and Concerns 241

longer available for use with one of the major programming tasks that every business per-
forms. However, despite the limitations you’ll run into, SOAP is still useful for database
application programming.
Don’t get me wrong. These issues are real and you need to consider them. For example, I’d Ch
never recommend sending mission-critical data across the Internet using SOAP because the
data might get lost in transit. Future specifications might eliminate this problem by using trans- 8
actions, but for now, you need to consider the problem as part of your development efforts.
There are however, thousands of non-mission critical database uses for SOAP. Consider the
lowly order entry system. Certainly, it’s no secret that your company accepts orders for mer-
chandise displayed in a catalog available to anyone. There isn’t any secret data to lose.
Crackers won’t compromise the order data if they do manage to get a peek. You can over-
come some of the obstacles of potential cracker activity by using an e-mail response system
that verifies order data with the client. Such systems are already in place and work well.

For the purposes of this book, the term cracker will always refer to an individual who
breaks into a system on an unauthorized basis. This includes any form of illegal activity
on the system. On the other hand, a hacker will refer to someone who performs low-
level system activities, including testing system security. In some cases, you need to
employ the services of a good hacker to test the security measures you have in place,
or suffer the consequences of a break-in. This book will use the term hacker to refer to
someone who performs these legal forms of service.

Of course, you do have to consider the sensitive data in an order entry system. You don’t want
to send the customer’s credit card number across the wire without protection or keep it on
your system for long. A customer might want to send credit card information using another
secure technique and simply reference that credit card using a code word on the order entry
page. We’ll investigate the issue of Web-based application security in the “Security Issues for
Web-Based Applications” section of Chapter 9, “Moving to Web-Based Applications.”

I realize that online purchasing systems already send credit card information across the
Internet using SSL. These companies store the data on a local server and allow you to
reference it from your machine. However, recent trade press stories suggest that it isn’t
a good idea to send this information across the wire, nor is it practical to store it
locally. Crackers have already invaded too many Web sites and stolen credit card infor-
mation. It’s safer, and more practical, to obtain this information using some secure
transfer method and store it only on a secure server behind a firewall for short time-
frames. The network administrator should purge old credit card information from inac-
tive clients on a regular basis.

Another important issue that many developers face when using SOAP for database manage-
ment is recordset size. Although it’s possible to transfer a moderate number of average-sized
records, you wouldn’t want to use SOAP to transfer a large number of records in a single
242 Chapter 8 Providing Remote Database Access

call. For example, one developer reported that he was able to transfer 300 records of about
1,200 bytes per record successfully. However, when he tried to move 1,000 records, SOAP
reported a timeout error. In addition, performance suffered when he was able to move a
large number of records. This problem doesn’t always occur—many developers write data-
base applications that move significant amounts of data in a single call—but it does happen
often enough that you need to consider it as part of your application design.
The main problem with transferring large numbers of records is that SOAP doesn’t provide
a direct representation of a record; it sends the data in an array of structures. The timeout
occurs when SOAP doesn’t receive all of the data in a predetermined time interval. (Some
SOAP toolkits allow you to set the interval using a special argument.) Variables such as line
noise make it difficult to predict how long a large download will take. SOAP toolkits also
process arrays differently, which means you’ll see different results by changing your toolkit.
For example, some developers reported better results using SOAP::Lite during the time of
this writing. (You can see SOAP::Lite messages related to databases at http://groups.
yahoo.com/group/soaplite/message/236.) Using SOAP:Lite allowed the developer to trans-
fer 11,000 records in as little as 25 seconds (which is still slower than a LAN). The toolkit
you choose will affect data transfer statistics.
No matter which toolkit you use, however, you’ll eventually run into time constraints.
Processing records using an Internet connection isn’t the same as using a LAN. The best
solution to the problem is to use server-side processing to restrict the number of records
you transfer to the client. Sending only the records you need to the client helps overcome
some of SOAP’s limitations.

Common Uses Based on SOAP Limitations

SOAP isn’t the end-all solution for database management. It has limitations that prevent it
from becoming the protocol of choice for heavy-duty applications. However, many classes of
light-duty applications work just fine with SOAP. The following list highlights some of the
important application types where SOAP works well despite limitations.
■ Surveys: Many companies rely on one-way data transfer to collect information from
customers. You can use SOAP to transfer data efficiently from a client site to the server
and then call upon a server-side component to process the data. Some detractors state
that HTML forms have been around for a long time and seem to fulfill every need.
However, companies are constantly creating new survey types, and SOAP represents
one way to add intelligence to the forms.
■ Forms: Companies run on forms. The fact that companies don’t make them out of
paper any more doesn’t change much. If an employee requires more supplies, he will
need to request them using some type of form. Processing these forms using SOAP will
reduce manager workload. In some cases, small purchases don’t require manager
approval, and the automation that SOAP provides allows the server-side application to
determine this automatically. A form application could also return a receipt to the user
telling him that the product is ordered. Another receipt might tell when the order actu-
ally ships from the warehouse.
 Developer Shortcuts for Remote Database Applications 243

■ Query: Queries represent one of the least potentially damaging forms of two-way
communication between client and server. A user makes a request; the server provides
the information in some form. SOAP provides the glue that allows the two to commu-
nicate. I’ve shown many examples of the query application in this book because users
Ch
need this kind of application most often. Even if your company doesn’t want to risk
data contamination by allowing users to enter data into the database, allowing the 8
users to query the database no matter where they are will still provide benefits in
employee performance.
■ Order Entry: If you have a sales force on the road, you know the convenience of remote
order entry. Customers are no longer willing to wait while the salesperson makes his
rounds to get an order started—they want their order entered today. However, smart
order-entry processing was problematic in the past because it normally relied on tech-
nologies such as DCOM. SOAP allows you to create order-entry systems that are more
reliable (at least when compared to DCOM on the Internet). Of course, you’ll want to
create safety procedures that protect user data, such as credit card numbers.
■ Reports: Besides forms, employees normally have a wealth of reports to fill out—
everything from their current status to the number of hours they worked in a week. In
the past, reports often waited until the employee came back from a trip, creating the
usual pile of papers that everyone hates to see. Using SOAP and SMTP together
allows an employee to transfer reports today that the server processes as it has time. As
with many applications that we’ve discussed in this list, you’ll need to buffer the
reports so the application doesn’t have remote access to the database. Unlike many
applications, the use of SMTP acts as a buffer, so you don’t need to build this support
into your application.

Developer Shortcuts for Remote

Database Applications
Converting your database application to use SOAP isn’t the same as creating a new applica-
tion. The amount of work that you put into the conversion depends on the complexity,
modularity, and overall workmanship of your application. A modular application usually
divides the middle-tier processing from the back end anyway, so everything is already in
place for the conversion. Sometimes, you can update the existing components and, in other
cases, you’ll create a small communication component to interact with SOAP.
Many developers find themselves overwhelmed with the task of updating an application to
use SOAP instead of another technology, such as CORBA. The thought is that every small
change at the front end of the project will create major changes in the back end or that the
project will run into compatibility nightmares from the beginning. Database applications are
especially daunting for all of the reasons we’ve discussed so far.
You can reduce the time required to update an application to SOAP or even create a new one.
The following tips provide a few ideas on how you can make your next project a little easier.
244 Chapter 8 Providing Remote Database Access

■ Test a simple case first: A friend recently reminded me about the test harnesses that
many manufactures use to check product components. For example, a washing machine
vendor might create a test harness to verify that the motor for that machine works as
anticipated long before the first washing machine appears on the scene. You can use this
same principle for your SOAP applications. Make sure you test a simple case of the
communication components before you invest time in creating the final user interface
or complex backend components. Data exchange is a major issue when your communi-
cation protocol relies on a text-based format.
■ Validate every database data type: Many database managers provide special data types to
allow efficient storage of bulky or unusual data. For example, most database managers
include currency and date types in native form. A database manager might include a pic-
ture or other unusual data type. Of course, SOAP doesn’t provide the functionality
required to transfer a picture data type, so you need to translate that data in some way. The
goal of validation is to ensure you can either transfer every database data type natively or
use a stored procedure to translate it before placing it on the wire. Creating generic rou-
tines now will ensure you don’t have to stop in mid-project to create the routines later.
■ Verify the toolkit’s capacity: Database applications tend to stress the capabilities of most
SOAP toolkits. You need to verify that the toolkit provides all of the functionality your
database application will require. For example, if you plan to transfer large amounts of
data with each transaction, you’ll need a toolkit that allows you to set the timeout value.
If your database includes a lot of unusual data types that your application will need to
use, then you’ll want a toolkit that supports user-defined types (UDTs). See the “Using
Complex Data Types” section of the chapter for details.
■ Choose a parsing technique early: The parsing technique you use depends on the kind
of data you need to transfer and the required transfer speed. Most XML parsers include
the ability to parse messages that rely on the Document Object Model (DOM) or
Simple API for XML (SAX). DOM is the World Wide Web Consortium (W3C) recog-
nized method for transferring data using XML and represents the best method to use if
compatibility is the major concern. DOM is also the best solution if you need to read
and write documents. However, many developers find DOM so complex to use that
they created SAX. SAX is more efficient than DOM and will allow you to create faster
applications. Unfortunately, SAX is read-only, so it limits the number of tasks you can
perform. A third alternative, JDOM, is supposed to provide the speed of SAX and the
flexibility of DOM, but it doesn’t have much of a market presence yet. Many SOAP
toolkits bury the method for using the parsing technology, so you’ll need to spend time
looking at how the documents call on the remote server.

The W3C finally recognized DOM Level 2 in November 2000. This is the standard that
adds support for Cascading Style Sheets (CSS), namespace support, a standard method
for combining objects, and a standard means to access and manipulate objects. In
short, the very basis of everything that SOAP represents is still in a state of flux. The
W3C is still considering other DOM changes that could ultimately affect how you work
 Using Complex Data Types 245

with SOAP. For example, you can expect DOM to support scalable vector graphics
(SVG) and mathematical markup language (MathML) in the future. The vendors
involved with DOM are currently working on the Level 3 specification.
Ch

■ Address client interoperability concerns: HTML is successful because you can use any 8
browser to view it—at least if the HTML coding follows the established standards. No one
has standardized SOAP yet, so there are interoperability problems. However, the SOAP
specification assures a certain level of compatibility between platforms. You need to ensure
that the toolkit you select will support all of your platforms and those of your clients.
■ Use standard database application development techniques: This book won’t tell you
how to create a database application. However, some developers change their tactics
every time they see a new technology because they think the old techniques will no
longer work. For the most part, developing a SOAP database application is no different
than developing one for your LAN. The only difference is the techniques you use to
transfer data from one place to another. Instead of instantiating components using
DCOM, you’ll use SOAP techniques.

Using Complex Data Types

Databases use a large number of data types. In fact, you might find that SOAP doesn’t sup-
port all of the data types for your DBMS. This means that you’ll face data conversion issues
that could affect the performance of your application to the point that handling thousands
of transactions per day becomes difficult.
The following sections address the issue of using complex data types within your applica-
tion. We’ll begin with an overview of the problem with working with complex data types.
The next section will show how various WSDL generators handle the problems of working
with complex data type. The last two sections show you a complex data type example.

Understanding the Problem

You need to consider two issues for data type conversions. The first is whether the unsup-
ported type is a complex type created from several simple types or whether it’s a new type.
You can create complex types in SOAP using a special coding technique. See http://www.
w3.org/TR/2001/PR-xmlschema-1-20010330/#element-complexType for details on the XML
definition for a custom data type. New types require a custom schema definition.

Remember that the basic difference between a simple type and a complex type is that
the complex type can carry elements in their content and include attributes. You can
define new simple or complex types, so it’s important to use the proper type for the
job at hand.
246 Chapter 8 Providing Remote Database Access

The issue is one of compatibility. SOAP supports complex types, even if you have to write
the code for them manually. (Few of the code generators that I’ve seen provide native sup-
port for complex types.) SOAP doesn’t support custom schemas. You can add the schema to
your message, but there’s no guarantee that the remote system will understand the new data
type. In short, custom data types allow you to convert data types to a form that a remote
system will understand, but new types are usually incompatible with other platforms.
Of course, you can always reduce the amount of data conversion you have to perform by
simulating the new data type using a complex data type. Consider the case where you have
to support a complex string. For example, SOAP doesn’t provide direct support for Pascal-
type strings. You could simulate this kind of string by defining a complex type consisting of
a number and a string. Here’s the Visual Basic version of such a definition.
Type ComplexStringType
StringLength As Integer
StringData As String
End Type

StringLength would hold the length of the string and make conversion easier. StringData
contains the string. Here’s what the XML would look like for such a type if you wanted to
use a formal definition:
<xsd:types>
<xsd:schema xmlns=”http://www.w3.org/2000/10/XMLSchema”
targetNamespace=”http://winserver/soapexamples/ComplexType/ComplexType.xsd” >
<xsd:element name=”ComplexStringType”>
<xsd:complexType>
<xsd:complexContent>
<xsd:restriction base=”xsd:anytype”>
<xsd:sequence>
<xsd:element name=”StringLength” type=”xsd:integer” />
<xsd:element name=”StringData” type=”xsd:string” />
</xsd:sequence>
</xsd:restriction>
</xsd:complexContent>
</xsd:complexType>
</xsd:element>
</xsd:schema>
</xsd:types>

However, you don’t always need to provide a formal definition of types in your application.
In fact, most developers will use the shortest form possible for convenience. Here is a per-
fectly acceptable alternative for most uses:
<xsd:types>
<xsd:schema xmlns=”http://www.w3.org/2000/10/XMLSchema”
targetNamespace=”http://winserver/soapexamples/ComplexType/ComplexType.xsd” >
<xsd:complexType name=”ComplexStringType”>
<xsd:sequence>
<xsd:element name=”StringLength” type=”xsd:integer” />
<xsd:element name=”StringData” type=”xsd:string” />
</xsd:sequence>
 Using Complex Data Types 247

</xsd:complexType>
</xsd:schema>
</xsd:types>

Let’s talk about the differences in the two definitions. You’ll remember from a previous dis-
cussion (see the “Attributes Versus Elements” section in Chapter 4, “Using SOAP to Create Ch
a Simple Application”) that you’ll normally want to use elements instead of attributes. Using 8
the name attribute for the <complexType> tag is acceptable, so we can get rid of the first
<element> tag. One of the two default content types for XML is <complexContent>, so we
can eliminate that tag as well. Finally, the <restriction> tag serves to restrict the acceptable
types for a complex type. Since we’ve used “anytype” in this case, the restriction doesn’t
exist and the tag is redundant. In fact, developers normally forgo using the <restriction>
tag unless they derive a new type from a base type. You can see the simplified form of com-
plex type definition in the specification (see the first example at http://www.w3.org/TR/
2001/PR-xmlschema-1-20010330/#Complex_Type_Definitions). However, it’s important to
remember that both formal and simplified definitions exist to allow better definition of
complex data types.
Of course, this is just one of many complex types that you might need to create for a data-
base application. You’ll find another write-up about complex data types in XML at
http://www.w3.org/TR/2001/REC-xmlschema-1-20010502/#element-complexType (or
http://www.w3.org/TR/xmlschema-0/#element-complexType). The example at this second
location shows how to use the simpleContent tag in place of the complexContent tag shown
in the example. You can use the simpleContent tag when you’re converting a simple type
without elements. In other words, the simpleContent type represents a method for doing
something like adding attributes to a simple type. You must use the complexContent tag
when converting types such as structures.

Understanding WSDL Generator Differences

Note that not all WSDL generators provide equal functionality when working with com-
plex types. The Microsoft SOAP Toolkit WSDLGen utility will create a WSDL file with
question marks that you need to remove. In addition, the modification of both the WSDL
and WSML files can prove troublesome. Using the psWSDL Wizard (described in
Appendix C) results in a single WSDL file that contains an entry for the user-defined type.
All you need to hand code, in this case, is the type definition, as shown in this section of
the chapter. In fact, the psWSDL Wizard even tells you where to place the type definition,
as shown in Figure 8.1.

The psWSDL example WSDL file and associated simple component form appear in the
\Chapter 08\ComplexString (Simple) directory of the source code available from the
Que Web site. You can find the source code at http://www.quepublishing.com.
The source code file also contains an IDL version of the component that works with
pocketSOAP. The two components perform the same task, but use different techniques
to do it.
248 Chapter 8 Providing Remote Database Access

Figure 8.1
Make sure you use the
correct WSDL genera-
tor for the job when
working with complex
data types.

The psWSDL Wizard also fixes another problem with the WSDL file. Visual Basic won’t
allow a developer to code user-defined types as ByVal. This means that when you gener-
ate the WSDL file using WSDLGen, any input arguments will also appear as outputs.
The psWSDL Wizard allows you to select the in/out status of each argument, which
means you get a better WSDL file output. However, as we’ll see later in this section,
you’ll still need to overcome dispatch problems to make the component work with
pocketSOAP.
Of course, creating a WSDL file doesn’t fix a more pressing problem with the Microsoft
SOAP Toolkit. This product doesn’t support UDTs using the high-level API that we’ve
used for many examples so far. You must write specially defined encoders and decoders
to do the job using the low-level API. Using the low-level API incurs a development
penalty that some developers won’t want to pay. You might find that converting UDTs
to classes is easier than creating a complex encoder and decoder when working with the
Microsoft SOAP Toolkit. You can access classes directly using the high-level API, but
again, you’re looking at additional coding effort if your application already uses UDTs.
Third-party products do support UDTs. For example, 4S4C (see Appendix D, “SOAP for
Visual C++ Developers,” for details) provides full UDT support. Figure 8.2 shows what
happens when you configure this product to access a DLL containing UDTs. As you can
see, 4S4C automatically generates the correct WSDL file for you. Unfortunately, this is
only the server side of the equation. You’d still have to create complex code for the client
side using the Microsoft SOAP Toolkit. That’s why this example relies on pocketSOAP
for the client.
 Using Complex Data Types 249

Figure 8.2
Some third-party prod-
ucts support UDTs
natively, but come
with other costs.
Ch

Describing the Interface

The server-side component is an amalgam of several techniques in this example. Visual Basic
alone won’t be able to create the component required for the complex data type because of
limitations in the language. In this case, you can’t control the in and out parameters as well
as you need to using native Visual Basic code. Visual Basic restricts your use of ByVal when
working with UDTs, which means that all parameters are treated as in/out parameters, an
unwanted condition when working with UDTs.
Getting around the Visual Basic limitation isn’t hard. All you need to do is create a type library
that explains how you want the parameters to work. The first task you need to perform is cre-
ating an Interface Definition Language (IDL) file similar to the one shown in Listing 8.1.

Listing 8.1 IDL File That Defines the Visual Basic Component Interface
// ComplexType.idl : IDL source for ComplexType.dll

[
uuid(5DA0875C-5F25-4B6F-977A-681B6A820B4B),
version(1.0),
helpstring(“ComplexType 1.0 Type Library”)
]

library ComplexType
{
importlib(“stdole2.tlb”);

typedef [uuid(EDD4BB08-52F8-402c-9F19-0AF3312D109A)]
struct ComplexStringType
{
short StringLength;
250 Chapter 8 Providing Remote Database Access

Listing 8.1 Continued

BSTR StringData;
} ComplexStringType;

[
object,
uuid(15E66EB2-693E-43D2-BAE2-F87BF8E0B05E),
oleautomation,
helpstring(“IComplexString Interface”),
pointer_default(unique)
]
interface ComplexString : IUnknown
{
[helpstring(“method GetComplexString”)]
HRESULT GetComplexString(
[in] ComplexStringType* InString,
[out, retval] BSTR* Result);
};
};

The first five lines of the IDL file define the globally unique identifier (GUID), version, and
help string for the type library as a whole. The help string is the text you see when you add
a reference to this type library to your Visual Basic project. The remainder of the IDL file
describes the library.

Using GUIDGen to Create GUIDs

If you haven’t had to create GUIDs before, you might wonder where they come from. You can create GUIDs in a
number of ways using Windows API calls, but this isn’t always useful because you need to know the GUID during
compile time. Visual Studio includes a utility named GUIDGen in the \Program Files\Microsoft Visual Studio\
Common\Tools directory. This tool doesn’t appear in the Visual Studio Tools menu found in the Start menu.
Open the GUIDGen utility; you’ll see a dialog box similar to the one shown in Figure 8.3. Notice that you can
create a variety of GUID formats. Normally, you’ll only use the registry format because it’s the most useful for
developers of all programming languages.
Creating a GUID is easy. Select Registry Format, and then click New GUID. Click Copy to copy the GUID to the
Clipboard and paste it into your IDL file. You can use these GUIDs anywhere you need to create a registry entry.
New complex data type applications require several GUIDs—three as a minimum for components of the type
shown in the example. For the purpose of this application, make sure you use the GUIDs shown in the source
code. Otherwise, you’ll break other code within the example and it won’t work as anticipated.

The library contains three code segments. The first is the description of the ComplexString
Type complex data type. As previously mentioned, this is a structure that includes a
StringLength and a StringData entry. This means the string isn’t null terminated but relies
on a length variable to indicate the end of the string. Note that you need to include a type-
def and a GUID for the ComplexStringType structure.
The second section is an object interface definition. You’ll require a GUID for each inter-
face in your IDL file. Listing 8.1 shows typical entries for working with Visual Basic,
although IDL supports many interface definition attributes. In this case, we’re telling Visual
Basic that the type library supports OLE automation and contains embedded pointers.
 Using Complex Data Types 251

Figure 8.3
The GUIDGen utility is
an important tool for
creating IDL files man-
ually.
Ch

The third section describes the ComplexString interface. This section normally contains a
list of the methods you want to export using the type library and is the main reason for
going this route. Note that you have complex control over the attributes for the
GetComplexString method. InString is an input-only parameter. Marking Result as both
output and retval (return value) means that you can set a variable equal to the output of this
method within Visual Basic. It’s also important to note that this third section requires no
GUID because it shares the GUID of the object as a whole.
After you create an IDL file, you need to compile it using the MIDL compiler. The source
code file available from the Que Web site contains a simple Make.CMD file that contains
the instructions necessary to compile the IDL file and register the type library. Here’s the
contents of the make file.
midl ComplexType.idl /newtlb /win32
regtlib ComplexType.tlb

The second line is important. Notice that you use RegTLib, not RegSvr32, to register a
type library. The RegTLib utility won’t appear on your server unless you installed Visual
Studio on it, so you need to copy the utility there. You must register both the component
and the type library on the server or the example application won’t work.

Creating the Server-Side Component

Now that we have a type library that describes an interface and the complex data type we’ll
use for this application, it’s time to create an implementation of that type library. The com-
ponent doesn’t do anything fancy in this case. All it does is check the length of the string
you’ve input against the length parameter and either pad or truncate the string as required.
I’m using dollar signs ($) so the padding shows.
Creating the component is relatively easy. All you need is a standard ActiveX DLL project;
then add a reference to the ComplexType 1.0 Type Library that we created in the previous
section. Listing 8.2 shows the source code for this example.
252 Chapter 8 Providing Remote Database Access

Listing 8.2 ComplexType Component Source Code

Implements ComplexString

Private Function ComplexString_GetComplexString( &_

InString As ComplexType.ComplexStringType) As String

Dim ResultString As String

‘Create a string of sufficient length.

ResultString = InString.StringData
If Len(ResultString) < InString.StringLength Then
Do While Len(ResultString) < InString.StringLength
ResultString = ResultString + “$”
Loop
Else
‘Truncate strings that are too long
If Len(ResultString) > InString.StringLength Then
ResultString = Left(ResultString, InString.StringLength)
End If
End If

‘Return it to the sender.

ComplexString_GetComplexString = ResultString

End Function

Remember that this component is an implementation of a type library, so the code begins by
showing which interface of that type library it implements. Visual Basic will check your code
to ensure you implement all methods contained within that interface. The method implemen-
tation is the ComplexString_GetComplexString() function, which returns the modified string.

Creating the Remote Client

The client application relies on a serialized interface with the server. This is similar to the
low-level interface employed by the Microsoft SOAP Toolkit. You’ll find that using the low-
level interface requires a little more code, but does provide the flexibility needed to use
UDTs in an application. Listing 8.3 shows the source code for the remote client.

Listing 8.3 Remote Client Source Code

Private Sub cmdTest_Click()

Dim SOAPEnv As Object

Dim Transport As PocketSOAP.HTTPTransport
Dim Param As Variant
Dim SOAPParam As Variant

‘Define the SOAP envelope.

Set SOAPEnv = CreateObject(“pocketSOAP.Envelope”)
SOAPEnv.MethodName = “GetComplexString”
SOAPEnv.URI = “http://winserver/soapexamples/ComplexType/”

‘Create a parameter to place within the envelope.

Set Param = SOAPEnv.CreateParameter(“InString”, “”, “SOAPStruct”)
 Using Complex Data Types 253

‘Initialize the parameter.

Set SOAPParam = CreateObject(“pocketSOAP.Param”)
SOAPParam.Init “StringLength”, txtLength.Text, “”
Param.Parameters.Append SOAPParam
Set SOAPParam = CreateObject(“pocketSOAP.Param”) Ch
SOAPParam.Init “StringData”, txtString.Text, “”
Param.Parameters.Append SOAPParam 8
‘Send the request and receive the data.
Set Transport = CreateObject(“pocketSOAP.HTTPTransport”)
Transport.Send “http://winserver/soapexamples/ComplexType/soap.asp”, &_
SOAPEnv.Serialize
SOAPEnv.parse Transport.Receive

‘Display the result.

lblResult.Caption = SOAPEnv.Parameters.Item(0).Value

End Sub

The code initially creates four variables. Notice how I’ve defined these variables. If you view
the variables in a debugger, it would seem that you could provide precise declarations, but
after much experimentation, I found these declarations work best. Almost every application
you create with pocketSOAP will require these four variables at a minimum.
The first step in creating the SOAP message is to create the envelope. Notice that
pocketSOAP requires two of the same parameters as the Microsoft toolkit: a method name
and a server location. You don’t need the name of a WSDL file because we’re not using one
for this example. Remember that this is a low-level interface example.
Creating one or more parameters comes next. This is one instance where using a complex
data type varies from simple data types. Normally you’ll create a variable that contains the
data you need. In this case, the parameter acts as a container for a structure containing two
variables. The next step of the process adds the two variables to the parameter. Note that
this is a three-step process that includes initializing a variable, adding the data to the appro-
priate structure member, and appending it to the envelope parameter.
After you create an envelope containing one or more method calls with the appropriate
parameters, you can serialize the data. The Transport variable sends the data to the
SOAP.ASP on the server. SOAP.ASP looks the required component up in CONFIG.XML,
instantiates the object, makes the required call, and sends the data back to the client. Notice
that the Send() method requires both a server-side location and the SOAP envelope we cre-
ated in the previous steps.
The Transport object receives data back from the server. However, the data is still in SOAP
message format, so you need to use the Parse method of the SOAP Envelope object to parse
it. Parsing the data separates the response from the rest of the information so that you can
display it. That’s the last step the code performs.
254 Chapter 8 Providing Remote Database Access

Testing the Complete Data Type Application

Unlike many of the applications in this book, this application relies on pocketSOAP
(http://www.pocketsoap.com/pocketsoap/) as a client because the Microsoft client doesn’t
provide UDT support. You need to download pocketSOAP and install it according to ven-
dor instructions. Likewise, the server relies on 4S4C (http://www.pocketsoap.com/4S4C/).
Appendix D provides a detailed discussion of this product. Make sure you install 4S4C on
the server using vendor instructions before you attempt to install the test component.
After you have the SOAP toolkits installed, copy the test component and type library to the
server. Make sure you register both the component (using RegSvr32) and the type library
(using RegTLib). I created a virtual directory for my Web server that contains SOAP.ASP,
SOAP_INC.ASP, SOAP-Error.HTM, CONFIG.XML, ComplexType.TLB, and
ComplexType.DLL. You must modify CONFIG.XML as shown here to allow 4S4C to find
the component. (Note that I split the second line of code over several lines to accommodate the
book width. You normally have to maintain a single long line between the two <mappings> tags.)
<mappings name=”demoService”>
<mapping name=”UDT_Demo” methodURI=”http://winserver/soapexamples/ComplexType/”
progid=”ComplexTypeSrv.ComplexStringImpl” iid=”{15E66EB2-693E-43D2-BAE2-
F87BF8E0B05E}”/>
</mappings>

Note that the file contains the name of the SOAP service, the mapped (friendly) name of the
component, the location of the component, the component’s program identifier, and the
interface GUID. The last two entries appear in the registry. You can also find them using
OLE View as we have done in other chapters. Make sure you list the interface GUID, not
the GUID for the component or the data structure.
Even if you get everything right, making this example work the first time can be tricky. You
know that a server configuration error exists if you have all of the required files in the virtual
directory and still receive an error message that the server can’t create the object. A number
of problems affect this example, such as the lack of rights to execute applications. After you
do configure everything correctly (including the Web server and server security), you can run
the application and see the output shown in Figure 8.4. Working with UDTs is doable, but
tricky. I found that after the application worked the first time, it worked reliably thereafter,
making this the kind of application that only presents problems when you first install it.

Figure 8.4
The output from the
ComplexType example
shows that you can
use UDTs with SOAP.
 Creating the Server-Side Component 255

Defining the SQL Server Database

The database for this example is relatively simple. It contains a list of tasks the user is sup-
posed to complete while on the road. Having a centralized task list allows an employer to
add new tasks to the list while the user is on the road. The user can update his list each time Ch
he calls into the company. Figure 8.5 shows the construction of this database. 8
Figure 8.5
A diagram of the sim-
ple database used in
this example.

This example relies on Microsoft SQL Server 7.0. The scripts found in the \Chapter
08\Data directory of the source code available from the Que Web site allow you to re-
create the database within SQL Server. You can find the source code at http://www.
quepublishing.com. The directory also contains a backup of the data so that you
don’t have to enter it manually.

As you can see, this database is relatively simple. It does include two datetime data fields so
that you can see the effects of using something other than text on your application. Because
datetime is a data type that SOAP supports (see Appendix A, “SOAP Data Types and Data
Type Conversions”), we should be able to treat it as a simple data type. In short, the applica-
tion shouldn’t require conversion routines. The Completed field is the only one that will
allow remote interaction by the user—all other fields are read-only.

Creating the Server-Side Component

This section of the chapter looks at the issues surrounding server-side components used for
database management. In some cases, you’ll find that the components you use for a SOAP
application are precisely the same components that you used for other applications created
for DCOM or CORBA. In other cases, you’ll find that you need to create new components
to handle data conversion or other issues that you might not have when working with binary
technologies.
We’ll also build a test component in this section. This component will interface directly
with the database, but won’t interface with the SOAP application. In fact, you’ll find a
DCOM application that can also interact with the database in the source code file available
from the Que Web site in the \Chapter 08\Task Checker directory. The point is that we’re
trying to separate the SOAP logic from the database management logic.
256 Chapter 8 Providing Remote Database Access

Many SOAP applications will work with the Microsoft SOAP Toolkit even if you don’t have
the latest version of their XML parser. However, database applications tend to stretch the
limits of the XML parser, so it’s a good idea to get the latest version to ensure good per-
formance. You can download the latest version (as of this writing) at http://msdn.
microsoft.com/downloads/default.asp?URL=/code/topic.asp?URL=/
msdn-files/028/000/072/topic.xml. Microsoft has also made some new tools
and samples available to SOAP developers at http://msdn.microsoft.com/code/
sample.asp?url=/msdn-files/027/001/457/msdncompositedoc.xml. Finally,
you can get the latest SOAP merge modules (those required for application distribution)
at http://www.microsoft.com/downloads/release.asp?ReleaseID=29556.

Tips for Working with Database Components

The biggest problem with most of the database projects that I update is that the original
developer used huge components that do everything but brush the administrator’s teeth.
This might seem like a good approach to some developers, but it makes it hard to work on
the project later. The best approach is to create multiple modules so you can add functional-
ity as needed. That’s why components are supposed to reduce the programmer’s work level;
they keep you from writing the same code repeatedly.
You can keep yourself out of hot water with the boss in a number of other ways. Database
components are some of the most complex components to write because the penalty for fail-
ure is so high and you’re always running into unforeseen problems. The following list pre-
sents some ways to decrease your workload, complete the project faster, force the
application to run faster, and maintain a high level of reliability.
■ Don’t overwhelm yourself with details. At least a few developers whom I talked with
have said that SOAP decreased their development time. One developer went so far as to
say that a project that would have normally required three months with DCOM,
required a mere month when working with SOAP. The secret all of these developers
had was not worrying too much about details such as security. They accepted the func-
tionality that SOAP had to offer at the outset of their project and saved time.
■ Use the low-level API only as required. Unless your application has a special need such
as using UDTs, you require the very last ounce of performance, or your application
won’t work with an IDispatch interface, always use the high-level API. Using the high-
level API saves considerable time and effort and it’s easier to debug.
■ Force existing components to do as much of the work as possible. You don’t want to
reinvent the wheel when working with SOAP. If you already have components in place,
make them perform the work. Even if you’re starting from scratch, you should consider
the desktop scenario. Create components in such a way that you can use more than one
access technique with small additions to the current component structure.
■ Be prepared to perform data conversions, even if it appears that you should be able to
use the native database types. Most SOAP toolkits have holes in their data type cover-
age. Some database types are incompatible with the XML specification. Likewise, some
programming languages implement standard data types or constructs in ways that
 Creating the Server-Side Component 257

SOAP can’t understand. All three of these issues conspire to make it difficult to transfer
anything but text consistently.
■ Always access the database in safe ways. Developers write many database components
with local access in mind. This means that you have to consider the method the compo-
Ch
nent normally uses to create connections and handle data. In some cases, the SOAP
component will have to perform additional work to ensure safe access. This includes 8
performing tasks such as closing the connection after each transaction.
■ Test all data exchange conditions fully. SOAP is a lot more sensitive to problems such as
NULL fields than the typical desktop application. You must resolve all data access, even if
it means creating an empty string to transfer across the wire. It’s also important to
remember that the XML parser will strip some control characters from the data stream.
For example, a carriage return/linefeed pair will become a linefeed—the XML parser
will always strip the carriage return from the data stream.
■ Use Base64 data encryption as needed to transfer binary data. Many databases provide
binary large object (BLOB) support. BLOBs never transfer across the wire when using
SOAP, so you must convert them to a format that SOAP can transport. Fortunately,
most SOAP toolkits provide conversion support as part of the package. For example,
the Microsoft SOAP Toolkit performs the conversion automatically if the XSD type in
the WSDL file is set to base64Binary.
■ Create a single result table to transfer to the client. A SOAP connection won’t support
multiple tables. You need to perform a join on the data to create a single result set that
the user can view. If the user needs another view of the data, the application will need to
request a new result set from the server.
■ Always use a static database connection. SOAP won’t allow dynamic connections at this
stage of the game. The best you can hope to achieve is a view of a static table and the
ability to send update information to the server. In many cases, it’s also important to
understand that updates might prove unreliable; therefore, they are risky.

Working with SQLXML

Microsoft has added an interesting new feature to SQL Server 2000: XML support. SQL
Server 2000 now accepts XML input and outputs XML to respond to client requests. Like
many other XML technologies, this one uses a specially formatted document. This is a
requirement because XML doesn’t specify the message format—it specifies the technique
used to create the message. SQLXML is pure XML. It helps you communicate directly with
SQL Server 2000 over an Internet connection.

Because SQLXML is such a new technology, information about it is a little scarce as of this
writing. Of course, magazines and book authors will soon take care of any holes in cover-
age for this new technology. In the mean time, you can find out more about SQLXML on a
few online sites. VBXML.COM (http://www.vbxml.com/people/speer/) includes an
overview of SQLXML and shows how to construct SQLXML messages. This site also hosts
258 Chapter 8 Providing Remote Database Access

a discussion group where you can talk about SQLXML. The New England SQL Server User
Group (http://www.nesql.com/downloads2.asp) has at least one presentation
about SQLXML on their Web site that includes a complete example of the technology at
work. Of course, you can always visit Microsoft’s site (http://www.microsoft.com/
sql/default.asp) and download an evaluation copy of SQL Server 2000 to see how
this technology works. Microsoft also provides an overview of SQL Server’s XML capability
(http://msdn.microsoft.com/library/techart/d51webapparch.htm) and
allows you to visit one of several SQL Server newsgroups (try microsoft.public.
sqlserver.xml or microsoft.public.sqlxml.viewmapper) to ask questions.

SQLXML works by creating an updategram. This is a special XML message that contains
instructions on what tasks the server should perform. The actions that you can perform
using SQLXML depend on the content of the schema for that database. The schema also
defines the database layout and other essential details. The schema contains namespaces
with desired actions. For example, the updg schema contains keywords such as before, after,
and sync to determine where new records are inserted into the database.
This technology also makes heavy use of stored procedures. The majority of the database
management work occurs when an XML message makes a request of a stored procedure and
passes it required data. This means you’re placing more of the processing burden on the
database manager, which is usually the least scalable part of a database application. It’s impor-
tant to consider performance when working with SQLXML—it probably isn’t the best choice
for exclusive access to large applications, but could work well for users on the road.
The SOAP connection occurs when you want to allow SQLXML to scale better. You can send
SOAP messages to local components that can monitor server load and upload the messages as
server load permits. The advantage of this method is that the components can be extremely
simple—you don’t need to translate the incoming XML data for the database manager.
Of course, the advantage works the other way as well. SQL Server can output XML that a
component can then wrap within an XML envelope without much effort. A component that
receives XML directly can work more efficiently than one that has to translate the data first.
Creating a SOAP envelope is relatively simple when compared to working with recordsets
directly.
Theoretically, using SQLXML would allow you to create the application without creating
multiple server-side components. Instead of creating both business logic and a buffer com-
ponent, you can create just the buffer component. However, although this solution will save
development time, it won’t scale as well as the traditional multi-component approach.

Using Multiple Server-Side Components

Generally, a database application employs several layers of business logic components that
massage the data before the database manager accepts it for storage. These components are
the most expensive part of the database application because they contain the most complex
logic. In short, if you have an existing application that you want to convert to SOAP, rewrit-
ing these components will represent a major part of the cost.
 Creating the Server-Side Component 259

Most SOAP applications will require a minimum of two server-side components. The first is
the database component. It provides standardized access to the database content. The sec-
ond is the SOAP component. It performs these three functions.
■ Converts the database data as needed Ch
Creates a buffer for the application to increase reliability and protect data
■
8
■ Creates the SOAP message using any techniques the toolkit supports

Depending on your server load, you might want to provide other components in the loop.
For example, it’s possible that the server load during the way will be so heavy that you’ll
need some form of offline storage to buffer the incoming data more than memory will allow.
In this case, you could add localized Queued Component (or the equivalent for the server
your company uses) to place messages on disk. The server can then pick up the messages as
processing capacity allows.
Using a secondary buffer can help reduce the risks associated with Internet application use.
It’s harder for crackers to create a denial of service (DOS) attack when the server processes
messages at its own pace. Yes, it’s true that the server’s hard drive will eventually fill with
messages the server can’t process fast enough, but using the secondary buffer buys you time
and allows the server to continue processing requests.
Some companies will require additional components if they use a layered security scheme.
The message travels through several layers of firewalls and other protection before the back
end server even sees it. You don’t need this level of security for a local connection, but it has
become a reality for Internet connections. Each server in the chain will require a set of
SOAP and server components. (This is where the SOAP actor feature comes into play—
multiple servers all perform part of the processing.)
The important consideration is to design the components for your applications carefully.
You’ll want to design the application in a way that allows you to “bolt on” SOAP support as
needed. Never use the database or business logic components on your system to support
SOAP. Otherwise, you might find that you have to redesign everything from scratch later.
There’s no guarantee that SOAP will provide everything your company needs; so, designing
to allow other forms of connection is essential. The example in this chapter uses DCOM as
the alternative connection, but you could easily use other protocols such as CORBA instead.

Generating the Code

The example uses a single server-side component to access the database. This database com-
ponent will allow you to transfer data in several ways, including DCOM. You’ll find a
DCOM client for this component in the \Chapter 08\Task Checker directory of the source
code file. Listing 8.4 shows what the code looks like. Notice that we’re performing standard
database tasks—nothing about the component indicates that we’ll eventually use it as part of
a SOAP application.
260 Chapter 8 Providing Remote Database Access

Listing 8.4 Server-Side Database Component Accessible from DCOM or SOAP

Public Function GetTasks(EmployeeNumber As String) As Variant
Dim TaskConn As ADODB.Connection ‘Database connection.
Dim TaskRecc As ADODB.Recordset ‘Recordset
Dim ConnStr As String ‘Database connection string.
Dim Result() As String ‘Array of result records.
Dim RecordCount As Integer ‘Total number of records in database.
Dim TotalCount As Integer ‘Total number of pertinent records.

‘Create a connection string.

ConnStr = “Provider=sqloledb;” & _
“Data Source=WinServer;” & _
“Initial Catalog=TaskList;” & _
“User Id=sa;Password=; “

‘Connect to the database.

Set TaskConn = New ADODB.Connection
TaskConn.Open ConnStr

‘Obtain the required recordset.

Set TaskRecc = New ADODB.Recordset
TaskRecc.Open “TaskTable”, _
TaskConn, _
adOpenStatic, _
adLockOptimistic, _
adCmdTable

‘Determine the number of pertinent records.

TotalCount = 0
For RecordCount = 0 To TaskRecc.RecordCount - 1
If TaskRecc!Employee = EmployeeNumber Then
TotalCount = TotalCount + 1
End If

TaskRecc.MoveNext
Next
ReDim Result(TotalCount - 1, 5)

‘Reset the variables.

TaskRecc.MoveFirst
RecordCount = 0
TotalCount = 0

‘Create an array of task record items.

For RecordCount = 0 To TaskRecc.RecordCount - 1

‘Verify the record item is for the current employee.

If TaskRecc!Employee = EmployeeNumber Then
Result(TotalCount, 1) = TaskRecc!Employee
Result(TotalCount, 2) = TaskRecc!Task_Name
If Not TaskRecc!Task_Description = “” Then
Result(TotalCount, 3) = TaskRecc!Task_Description
Else
Result(TotalCount, 3) = “No Task Description Available”
End If
Result(TotalCount, 4) = TaskRecc!Date_Assigned
If Not TaskRecc!Completed = “” Then
 Creating the Server-Side Component 261

Listing 8.4 Continued

Result(TotalCount, 5) = TaskRecc!Completed
Else
Result(TotalCount, 5) = “Not Completed”
End If Ch

TotalCount = TotalCount + 1 8
End If

‘Move to the next record.

TaskRecc.MoveNext
Next

‘Return the task array to the caller.

GetTasks = Result

‘Close the recordset and database.

TaskRecc.Close
TaskConn.Close

‘Clear the variables.

Set TaskRecc = Nothing
Set TaskConn = Nothing
End Function

Sub SetComplete(EmployeeID As String, TaskName As String, Assigned As Date)

Dim TaskConn As ADODB.Connection ‘Database connection.
Dim TaskRecc As ADODB.Recordset ‘Recordset
Dim ConnStr As String ‘Database connection string.
Dim SearchStr As String ‘Database search string.

‘Create a connection string.

ConnStr = “Provider=sqloledb;” & _
“Data Source=WinServer;” & _
“Initial Catalog=TaskList;” & _
“User Id=sa;Password=; “

‘Connect to the database.

Set TaskConn = New ADODB.Connection
TaskConn.Open ConnStr

‘Create a search string.

SearchStr = “Select * From TaskTable Where “ & _
“Employee = ‘“ & EmployeeID & _
“‘ and Task_Name = ‘“ & TaskName & _
“‘ and Date_Assigned = ‘“ & Assigned & “‘“

‘Obtain the required recordset.

Set TaskRecc = New ADODB.Recordset
TaskRecc.Open SearchStr, _
TaskConn, _
adOpenStatic, _
adLockOptimistic, _
adCmdText

‘Set the completed field to the current date.

TaskRecc!Completed = Date
262 Chapter 8 Providing Remote Database Access

Listing 8.4 Continued

‘Update the record.
TaskRecc.Update

‘Close the recordset and database.

TaskRecc.Close
TaskConn.Close

‘Clear the variables.

Set TaskRecc = Nothing
Set TaskConn = Nothing
End Sub

As you can see, this component has two methods. The first, GetTasks(), returns an array of
tasks for the employee making the request. The second, SetComplete(), will place a date in
the Completed field for the user. Given the simple nature of the database we’re using, these
two functions provide nearly full control (at least as much as you would want a remote user
to have over your system). Notice the lack of destructive commands, such as the ability to
erase a record.
Both methods begin by creating a connection to the database. This includes creating a con-
nection object and a recordset. The GetTasks() method obtains the entire table for the
recordset, while the SetComplete() method returns a single record based on the user’s cur-
rent position within the task list. It’s possible to increase the efficiency of the GetTasks()
method by using a more selective record-retrieval method, but the method shown works
fine for smaller databases, such as the one in the example.
The GetTasks() method performs several processing steps on the recordset. First, it deter-
mines how many records the database contains for the user in question. It uses the result of
this counting operation to redim a dynamic array. After the component knows how big to
make the array, it resets all of the variables and fills the array with data. Notice the checks
for NULL values for fields that allow NULLs in the database. Failure to check for NULLs will
cause the application to fail.

Always use dynamic arrays when working with databases because you don’t know in
advance how many records the array will need to store. In addition, Visual Basic will
display an error if you try to copy a static array to a variant. This is a known issue and
Microsoft provides suggestions for working around it in the Visual Studio documenta-
tion. Unfortunately, dynamic arrays are difficult to transfer using SOAP. We’ll see later
in this chapter how this affects the application. Arrays represent one of the major
trade-offs of using SOAP for your application; Visual Basic and SOAP have completely
different requirements in this arena.

The SetComplete() method has life a little easier. After it obtains a copy of the recordset, it
changes the date in the Completed field and updates the record. Theoretically, the change
should take place immediately. However, even if the change does occur immediately, the
constraints placed on the application by SOAP means that at least some users would have
 Creating a Middle-Tier Component 263

outdated information loaded on their machines. If providing updated information is a prior-

ity for your application, you’ll need to add some type of timed update to the application.
Events might not work very well, in this case, because SOAP applications don’t maintain a
consistent connection with the server, something that events rely on to do their job.
Ch
Both components end by closing the connection and recordset. Keeping track of
resource usage with SOAP applications is essential because they tend to request more 8
resources than a desktop application in many cases. For example, SOAP applications
require a new connection every time they make a request. Therefore, you need to check
code for proper database closures and object releases. Although the system will compen-
sate for some level of negligence in this area, failure to release resources will eventually
eat away at application performance and could cause the server to crash when resources
are exhausted.

Creating a Middle-Tier Component

You might ask what could be so different about creating the middle-tier component than
simply moving data from the server to a SOAP message. The problem is that SOAP doesn’t
handle odd data formats very well. The server-side component in Listing 8.4 outputs an
array using a variant. Theoretically, SOAP will handle an array without problems. In fact, if
we create a component that simply passes the data along as shown in Listing 8.5, the various
WSDL generators will generate the correct output (except for 4S4C in this case). Figure 8.6
shows typical output.

Listing 8.5 Just Passing the Data to SOAP as Shown Here Won’t Work
Public Function GetTasks(EmployeeNumber As String) As Variant
‘Create a task database interface object.
Dim GetTask As DataAccess.TaskTable
Set GetTask = New DataAccess.TaskTable

‘Get a list of current tasks.

TaskTable = GetTask.GetTasks(EmployeeNumber)

‘Pass the list to the client.

GetTasks = TaskTable

‘Clean up the objects.

Set GetTask = Nothing
End Function

Notice that the code looks fine, as does the WSDL output. The WSDL Generator utility
detects the date variable for the SetComplete() method and handles the variant as an
xsd:anyType. The problem is the line highlighted in Figure 8.6. If you run this code as is,
you’ll receive the error message shown in Figure 8.7 from the client.
Notice that the error message states that the SoapMapper was unable to work with the
anyType variable. Fixing this problem means performing one of three tasks. You can hand
edit the WSDL file to replace the anyType value with an array, create an IDL file, or change
264 Chapter 8 Providing Remote Database Access

the data into a format that’s more palatable for SOAP. (See the complex data type example
earlier in the chapter for an example of an IDL file.)

Figure 8.6
The WSDL output
looks correct, but
won’t generate a good
result.

Figure 8.7
SOAP will generate an
error message for the
client when working
with certain types of
variants.

Many professionals suggest you avoid using the variant type for SOAP applications.
XML does support this data type, and so does SOAP. The problem is that the SOAP
mapper often gets confused when working with variant data and assumes that it can’t
place the requested data within the variable. Arrays are simple constructs to place
within a variant, yet most developers report they can’t use them with any SOAP toolkit.
Trying to map complex data such as images into a variant will fail more often than not.
In most cases, you’ll want to use Base64 encoding for complex data and avoid using
variants even if they are supported by your SOAP toolkit.
 Creating a Middle-Tier Component 265

We’ll use the third technique in this section. The reason that I chose this method is that it
takes the least time to implement and promises the greatest probability of success on the
first attempt. Listing 8.6 shows a different version of the SOAP component that modifies
the data into a string that we’ll parse on the client end.
Ch

Listing 8.6 A SOAP Component That Works 8

Public Function GetTasks(ByVal EmployeeNumber As String) As String
Dim RecordCount As Integer
Dim TaskString As String

‘Create a task database interface object.

Dim GetTask As DataAccess.TaskTable
Set GetTask = New DataAccess.TaskTable

‘Get a list of current tasks.

TaskTable = GetTask.GetTasks(EmployeeNumber)

‘Convert the array into a string that we can parse.

TaskString = “”
For RecordCount = 0 To UBound(TaskTable)
TaskString = TaskString & TaskTable(RecordCount, 1) & vbCrLf
TaskString = TaskString & TaskTable(RecordCount, 2) & vbCrLf
TaskString = TaskString & TaskTable(RecordCount, 3) & vbCrLf
TaskString = TaskString & TaskTable(RecordCount, 4) & vbCrLf
TaskString = TaskString & TaskTable(RecordCount, 5) & vbCrLf
Next

‘Pass the list to the client.

GetTasks = TaskString

‘Clean up the objects.

Set GetTask = Nothing
End Function

Sub SetComplete(ByVal EmployeeID As String, _

ByVal TaskName As String, _
ByVal Assigned As Date)
‘Create a task database interface object.
Dim GetTask As DataAccess.TaskTable
Set GetTask = New DataAccess.TaskTable

‘Pass the completed field data to the server.

GetTask.SetComplete EmployeeID, TaskName, Assigned

‘Clean up the objects.

Set GetTask = Nothing
End Sub

This version of the code generates a much smaller WSDL file, as shown in Figure 8.8.
Notice that we’re using a string to transfer the data now. The string will only incur a small
increase in data size and shouldn’t affect performance.
266 Chapter 8 Providing Remote Database Access

Figure 8.8
Using the alternative
code produces a
smaller WSDL file that
relies on simple data
types.

Let’s discuss the SOAP component code in a little more detail. Notice that the component
contains two methods, just like the database component, and that it uses the same method
names. A few developers would say this causes confusion, but you want to maintain some
level of consistency at the client end of the SOAP application picture. Both components use
the same calling convention. The SOAP component returns a string instead of a variant for
the GetTasks() method call.
Both methods begin by creating an instance of the database component we discussed in the
previous section. In this regard, they act just as a client-side application would. Comparing
this component code with the DCOM client example on the Web site will show some amaz-
ing similarities.
The GetTasks() method retrieves the array of records from the database component. It
uses a loop to convert the individual records into a single string for transfer to the
client. The example code uses a carriage return/linefeed combination between each field
entry. The XML parser will strip off all of the carriage returns, but the linefeeds will
remain in place. You could use any special character to provide separations in the data
for later parsing. However, using a carriage return/linefeed pair makes it easier to read
the data using utilities such as tcpTrace. It’s also unlikely that you’ll ever run into either
a carriage return or a linefeed within the database data, so this pair represents a reliable
parsing character.
The SetComplete() method simply passes the data from the client to the server. In this case,
we need to worry about the date variable. However, most SOAP toolkits provide good date
and time type support, so it shouldn’t be too much of an issue.
 Creating the Client-Side Application 267

Creating the Client-Side Application

The client application requires more code than any other application in this book. No mat-
ter how you get data from the server to the data, it always requires some amount of inter-
pretation and formatting. SOAP adds to this burden by forcing you to convert data in many Ch
situations where you wouldn’t need to with a desktop application. Listing 8.7 shows the util- 8
ity methods for this example.

Listing 8.7 Utility Routines for the Client Application

Private Sub cmdQuit_Click()
‘Exit the program.
End
End Sub

Public Sub DisplayText()

‘Display the fields within the current record.
txtEmployeeID.Text = TaskTable(RecordNum, 1)
lblTitle.Caption = TaskTable(RecordNum, 2)
lblDescription.Caption = TaskTable(RecordNum, 3)
lblAssigned.Caption = TaskTable(RecordNum, 4)
lblComplete.Caption = TaskTable(RecordNum, 5)
End Sub

Public Sub ParseData()

Dim TextPos As Integer
Dim RecordCount As Integer

‘Calculate the number of records.

TextPos = 1
RecordCount = 0
Do
RecordCount = RecordCount + 1
TextPos = TextPos + 1
TextPos = InStr(TextPos, TaskString, Chr(10))
Loop While TextPos > 0
RecordCount = RecordCount / 5

‘Dimension the table to hold the number of records.

ReDim TaskTable(RecordCount - 1, 5)

‘Parse the data in a loop.

RecordCount = 0
Do
TaskTable(RecordCount, 1) = Left(TaskString, _
InStr(TaskString, _
Chr(10)) - 1)
TaskString = Right(TaskString, _
Len(TaskString) - InStr(TaskString, _
Chr(10)))
TaskTable(RecordCount, 2) = Left(TaskString, _
InStr(TaskString, _
Chr(10)) - 1)
268 Chapter 8 Providing Remote Database Access

Listing 8.7 Continued

TaskString = Right(TaskString, _
Len(TaskString) - InStr(TaskString, _
Chr(10)))
TaskTable(RecordCount, 3) = Left(TaskString, _
InStr(TaskString, _
Chr(10)) - 1)
TaskString = Right(TaskString, _
Len(TaskString) - InStr(TaskString, _
Chr(10)))
TaskTable(RecordCount, 4) = Left(TaskString, _
InStr(TaskString, _
Chr(10)) - 1)
TaskString = Right(TaskString, _
Len(TaskString) - InStr(TaskString, _
Chr(10)))
TaskTable(RecordCount, 5) = Left(TaskString, _
InStr(TaskString, _
Chr(10)) - 1)
TaskString = Right(TaskString, _
Len(TaskString) - InStr(TaskString, _
Chr(10)))
RecordCount = RecordCount + 1
Loop While Len(TaskString) > 0
End Sub

The application uses a separate DisplayText() function to save development time. As we’ll
see in the next section, the application needs to display data from several different locations.
There isn’t anything too special about this routine except that the RecordNum variable tracks
the current record within the TaskTable array.
The ParseData() function requires a little more explanation. The code begins by calculating
the number of records the application received. The SOAP component in the previous sec-
tion could have passed the number of records as part of its output, but it’s better to calculate
this number at the client so that you can account for any lost records. Lost data is always
possible when using an Internet connection; so, calculating the number of records at each
endpoint is essential.
After the application knows the number of records, it can ReDim the TaskTable array and
begin placing data within it. Notice that the parsing routine uses standard text manipulation
functions to locate the position of linefeeds within the data stream. As previously mentioned,
the XML parser strips any carriage returns from the data stream, so you won’t find any.
It’s important to parse the data carefully. You not only have presentation concerns to worry
about, but interaction with the database to consider. A stray control character will contami-
nate the data within the array. Passing this contaminated data back to the server (as we will
with the SetComplete() method) results in searches that fail and other problems. Given the
inability to provide robust error handling with SOAP, you’ll spend a lot of time debugging
applications to find errors in parsing. In short, test your parsing routine thoroughly before
placing it in the application.
 Creating the Client-Side Application 269

One of the biggest issues in creating a database application is safely moving the data from
the server to the client. The client requires some type of mechanism to control precisely
which data the server returns. This is especially important with SOAP application because
you’ll likely move data across connections that are slower than the ones provided by the typ-
Ch
ical network. Listing 8.8 shows the data exchange methods for this example.
8
Listing 8.8 Data Exchange Routines for the Client Application
Dim TaskTable() As String ‘Task Record Array
Dim TaskString As String ‘Unparsed Task Record String
Dim RecordNum As Integer ‘Current Record Number

Private Sub cmdCompleted_Click()

‘Create the SOAP client.
Dim Client As SoapClient
Dim Result As String
Dim ErrorMessage As String

‘Set up an error handler.

On Error GoTo ErrorHandler

‘Create the connection.

Set Client = New SoapClient
Client.mssoapinit _
“http://localhost/soapexamples/TaskList/SOAPData.WSDL”, _
“SOAPData”, _
“TaskTableSoapPort”

‘Set the completed field.

Client.SetComplete txtEmployeeID.Text, _
lblTitle.Caption, _
lblAssigned.Caption
Set Client = Nothing

‘Display the current task again.

Set Client = New SoapClient
Client.mssoapinit _
“http://localhost/soapexamples/TaskList/SOAPData.WSDL”, _
“SOAPData”, _
“TaskTableSoapPort”
TaskString = Client.GetTasks(txtEmployeeID.Text)
ParseData
DisplayText
Set Client = Nothing

‘We’re finished, so exit.

Exit Sub

‘Display a message when an error occurs.

ErrorHandler:

ErrorMessage = “Fault Code: “ + Client.faultcode + vbCrLf + vbCrLf + _

“Fault String: “ + Client.faultstring + vbCrLf + vbCrLf + _
“Fault Actor: “ + Client.faultactor + vbCrLf + vbCrLf + _
“Detail: “ + Client.detail
270 Chapter 8 Providing Remote Database Access

Listing 8.8 Continued

MsgBox ErrorMessage, vbExclamation
End Sub

Private Sub cmdTask_Click()

‘Create the SOAP client.
Dim Client As SoapClient
Dim Result As String
Dim ErrorMessage As String

‘Set up an error handler.

On Error GoTo ErrorHandler

‘Create the connection.

Set Client = New SoapClient
Client.mssoapinit _
“http://localhost/soapexamples/TaskList/SOAPData.WSDL”, _
“SOAPData”, _
“TaskTableSoapPort”

‘Get a list of current tasks.

TaskString = Client.GetTasks(txtEmployeeID.Text)
ParseData

‘Set the current record number.

RecordNum = 0

‘If there is more than one record, enable

‘the Next button.
If UBound(TaskTable) > 0 Then
cmdTask.Enabled = False
cmdTask.Visible = False
cmdNext.Enabled = True
cmdNext.Visible = True
cmdNext.SetFocus
cmdPrevious.Enabled = True
cmdPrevious.Visible = True
End If

‘Enable the Completed button.

cmdCompleted.Enabled = True
cmdCompleted.Visible = True

‘Display the current task.

DisplayText
Set Client = Nothing

‘We’re finished, so exit.

Exit Sub

‘Display a message when an error occurs.

ErrorHandler:

ErrorMessage = “Fault Code: “ + Client.faultcode + vbCrLf + vbCrLf + _

“Fault String: “ + Client.faultstring + vbCrLf + vbCrLf + _
“Fault Actor: “ + Client.faultactor + vbCrLf + vbCrLf + _
 Creating the Client-Side Application 271

Listing 8.8 Continued

“Detail: “ + Client.detail
MsgBox ErrorMessage, vbExclamation
End Sub
Ch

Both the cmdCompleted_Click() and the cmdTask_Click() methods begin the same way as 8
many other examples in the book. They create a SOAP client and set up error handling.
Notice that they both use the same access technique. The client connection controls access
to the component as a whole, not to an individual method within the component.
One point of interest is that the cmdCompleted_Click() method contains two invocations of
the SoapClient. After many hours of experimentation, I found that SOAP normally drops
the connection between client and server long before you can make two method calls to the
remote server. The dropped connection produces an error message similar to the one shown
in Figure 8.9. As a result, you must re-create the SOAP connection for every method call,
unlike desktop applications where the connection remains intact. This requirement empha-
sizes the fact that SOAP provides a one-way, one-time connection to the server. Note that
both methods provide extended error information—you’ll need this additional information
to debug most database applications.

Figure 8.9
Dropped HTML con-
nections produce an
error message similar
to the one shown
here.

The cmdCompleted_Click() method calls the SetComplete() method first. Notice that the
three input values come from the client application dialog box. This is the best place to get
the information because it shows the current task. Calling SetComplete() modifies the data-
base, but it doesn’t update the display. The cmdCompleted_Click() method also calls on
GetTasks() to update the TaskString and ParseData() to update the TaskList array. Finally,
it displays the updated information onscreen.
The cmdTask_Click() method displays the task data using the same techniques as
cmdCompleted_Click(). cmdTask_Click() also enables various buttons depending on how
many task records the application retrieves. The application always enables the
cmdCompleted button after it retrieves the first record.
The last two methods for this example appear in Listing 8.9. Both methods move from
record to record by changing the RecordNum value and displaying the new record on screen.
Both methods also check for invalid RecordNum values. They display an appropriate message
when the user is already at the end of the list.
272 Chapter 8 Providing Remote Database Access

Listing 8.9 Movement Routines for the Client Application

Private Sub cmdNext_Click()
‘Get the next record.
If RecordNum < UBound(TaskTable) Then
RecordNum = RecordNum + 1
DisplayText
Else
MsgBox “Already at last task!”, _
vbExclamation Or vbOKOnly, _
“Record Position”
End If
End Sub

Private Sub cmdPrevious_Click()

‘Get the previous record.
If RecordNum > 0 Then
RecordNum = RecordNum - 1
DisplayText
Else
MsgBox “Already at first task!”, _
vbExclamation Or vbOKOnly, _
“Record Position”
End If
End Sub

Testing the Application

Depending on your company’s requirements, database applications normally require several
levels of testing. For this example, I performed three levels of testing:
■ Local testing using a data editing application
■ Remote testing using DCOM to simulate a desktop application
■ Remote testing using SOAP to simulate client access across the Internet

Local testing is important because it allows you to test the database implementation without
incurring the cost of remote transactions. I used the simple data editor shown in Figure 8.10
for this example. This data editor also comes in handy for testing new records and resetting
the database after a test.
Make sure you test everything the database has to offer during local testing. This means
testing any stored procedures, checking views, and trying the various indexes. It’s relatively
easy to change the configuration of the database during the early design phases. Creating a
local test program allows you to check the database thoroughly to ensure your design has no
holes before you begin remote testing.
Using DCOM or some other binary technology to perform remote testing isn’t a require-
ment. However, you’ll find that you can test the server-side components better if you use
this intermediate strategy. Using DCOM allows you to test the server-side database compo-
nents separately from the SOAP components. Performing this individual level of testing
 Testing the Application 273

reduces the number of places you have to look for errors. Figure 8.11 shows the client appli-
cation interface used for both the DCOM and the SOAP versions.

Figure 8.10
Local tests are essen- Ch
tial if you want to
check the database for 8
errors.

Figure 8.11
The Task Checker user
interface presents the
task list information in
a non-editable format.

You won’t have access to your component after testing it with SOAP. The system locks
the DLL file because the DLL is still loaded in memory. Use the IISRESET command-line
utility to stop and start IIS to unload the component. Using this technique will allow you
to update the DLL as needed for testing purposes without having to restart the machine.
(Unfortunately, simply starting and stopping IIS won’t accomplish anything.) You can
achieve the same effect by stopping and starting IIS from the Services MMC snap-in. A
third option is to use net stop iisadmin /y to stop IIS, then net start w3svc
to restart it. Some developers place these two commands in a batch file so they only
have to remember one command. The IISRESET command is the fastest method, how-
ever, so it’s probably the best option for developers with local system access.
274 Chapter 8 Providing Remote Database Access

After you begin testing the SOAP version of the application, you need to check the SOAP
component and how it interacts with the database component on the server. You perform
most of the business logic testing as part of creating a DCOM client. The SOAP application
is a “bolt on” to the existing DCOM (or other binary protocol) application.
For many developers, performing all of these levels of testing is out of the question. They
find that their testing time is short enough as it is without adding complexity to the testing
process. The problem with testing a complex database application after you integrate all of
the components is that you’ll find it’s difficult to locate problem components. The reason
for all of these levels of testing is to simplify the testing process, making it easier to locate
bugs and fix them.

Quick Fixes for Remote Database Applications

Database applications are becoming more complex every day. As companies create new
sources of information and combine existing ones, database developers discover the com-
plexity of the applications they create increases exponentially. In addition, companies are
more global than they were in the past. The issue of 24/7 up-time is becoming the norm,
and many developers find that it places additional requirements on their applications. No
longer can developers rely on a maintenance period to compress data, re-create indexes, and
generally clean up the application environment.

If you’re worried about the number of interdependent specifications and standards that
your SOAP applications rely upon, you’re not alone. Many developers have noticed that
XML in general is becoming a mass of interdependent specifications, and organizations
such as W3C seem determined to output more as quickly as possible. I was recently
online and noticed that one person had put all of the pieces together into a comprehen-
sive (if not comprehendible) list. You can read more about the problems of interdepen-
dent standards at http://www.xml.com/pub/a/2001/02/21/deviant.html.
Although this represents the opinion of one person, it also provides a good view of the
problems most developers are facing with this ever-expanding technology.

Time is the enemy of the database developer. When a company experiences down time for
other application types, it might affect just a few people. The requirement for getting the
application fixed isn’t as great as when a broken database application affects a majority of the
company. That’s why this section is titled “Quick Fixes for Remote Database Applications.”
Not only do database applications present special challenges, but also few people are willing
to wait for you to fix them.

Many issues can affect the reliability and usability of your application when using SOAP.
For example, some developers are concerned that Microsoft programmed many parts of
the Microsoft SOAP Toolkit using a single threaded apartment (STA). This can cause
problems when developers call the various objects, such as HttpConnector, from a
 Quick Fixes for Remote Database Applications 275

multi-threaded apartment (MTA) application. The bottom line is that you need to under-
stand how the developer created your SOAP toolkit to ensure you can use it in all
required application scenarios.
Ch

We’ve already discussed many potential problem areas as part of the examples in this chap- 8
ter. However, there are other problems that you should consider during the design, develop-
ment, and debugging stages of your application. The following sections discuss some
database scenarios that are specific to SOAP.

Loss of Connection
The Internet presents many ways to lose a connection. Line noise and other hazards prevent
even the best connection from working all of the time. In addition, you have to consider the
effects of loading because most ISPs overbook their line capacity. In other words, you’ll run
into situations where there isn’t enough bandwidth bandwidth to support a database applica-
tion fully because the ISP has sold part of the anticipated bandwidth to another customer.
Some ISPs do sell guaranteed capacity at a huge markup. Having more than one ISP is
more likely to provide the level of redundancy you require for a database application.
You also need to consider SOAP-specific issues. Most SOAP toolkits provide a watchdog
feature. This feature looks for open connections without activity. If the timeout value
expires, the connection is lost and the client receives an error message. Most of these toolk-
its also allow you to set the timeout value for the connection. Some developers might con-
sider setting the timeout to a high value to ensure that the connection never fails due to
internal processing. However, using a high value sets your application up for a “hung” state
if the connection really does fail.
Setting the timeout value requires experimentation in a real world setup with your particular
application. Several factors affect the setting, including the number of users, type of connec-
tion, amount of data, and level of processing. An order entry system might require a longer
timeout than a catalog search application simply because the order entry system experiences
a higher level of activity and passes more data between the client and server.
SOAP connections also fail when the unexpected happens. You might find that certain client
and server combinations will disconnect, rather than display an error message. This hap-
pened with several of the test applications in the book, resulting in code and tool changes.
In theory, SOAP applications should communicate errors, but the reality is that vendors
don’t always anticipate error-handling scenarios correctly, and the client or server side of the
application crashes.

Odd Data Entry Errors

Any time you convert data from one format to another, you risk loss of content or misinter-
pretation of data. The fact that SOAP forces at least two data conversions for every data
transfer means that there’s a high probability you’ll see conversion errors sometime during
your coding experience. Database applications often require more than just the two forced
276 Chapter 8 Providing Remote Database Access

conversions. For example, the second program in this chapter requires a parsing mechanism
at both ends of the application to convert an array to a string and back. In short, the sample
application requires five total conversions: database to array, array to string, string to XML,
XML to string, and string to array.
The fastest way to correct data entry problems is to create the data entry components and
create a test suite to check the output of each component individually. If you can reduce the
code to a black box of component implementations, integration testing becomes much eas-
ier. Make sure you check all types of data input a component might receive, including those
that the component should never have to handle. Always test the component’s ability to han-
dle NULL input and empty values. It’s especially important with strings to check that the data
you input is the data that the component outputs as well. One example in the book caused
many problems until I discovered the output from a component was one character (a space
no less) too short. Errors like this are difficult to locate and fix.
Always create a test database as part of the testing process. I create a loopback test applica-
tion that receives output from a database through a series of components, turns the data
around, and sends it right back to the database using the same components. The data you
see in the database at the end of the test should precisely match the original data. Don’t for-
get that you can use applications such as tcpTrace in these scenarios to log the flow of data.
This allows you to view the data later for discrepancies in handling by specific components.

Performance Issues
Early adopters of SOAP are already complaining that database performance is much worse
than when using alternatives such as DCOM and CORBA. The problem is that a database
application transfers fields of information within individual records. The data can grow by as
much as 10 times its normal size due to the tags that XML uses to define the schema.
Obviously, bandwidth begins to play a major role in performance at this point.
Therefore, the question of performance becomes one of reducing data growth in the XML
portion of the message. One way to do this is to use smaller tags. However, generating a
message by hand to ensure the tags stay small is impossible. The automated way to do this is
to ensure you use smaller variable names for component method arguments. Make sure the
name is descriptive, yet short enough to keep the resulting tags small when automatically
generated by an application.
Another potential fix for this problem is to hand edit the WSDL file used to transfer the
data. The problem with this approach is that even a small editing error can keep your appli-
cation from running at all. However, careful hand editing will help reduce performance-
related problems in your application.
I had previously mentioned that one of the testing phases you should use for a database
application is to check the server-side components using a DCOM or CORBA client. Many
vendors provide utilities to stress test components under such conditions. In some cases,
developers avoid stress-testing their application because it’s time consuming and the perfor-
mance benefits gained are small compared to the time required to implement the changes.
 Addressing Transaction Issues 277

SOAP applications are different—you need every advantage you can get to reduce the size
and processing time of the message.
A few developers write their own XML parser to gain a speed advantage. Although writing an
XML parser isn’t a daunting task, it could become time consuming. However, finding a fast Ch
XML parser for your system should be a top priority. It isn’t hard to imagine that magazines
will begin to review the merits of various XML parsers by the time you read this chapter. 8
Server Is Busy or Missing Objects
If you receive a message that the server is busy, that the SOAP object is missing, or that the
server experienced an unknown error, don’t feel alone—these are frequent problems for
everyone. However, these problems normally occur because something simple is wrong with
the server setup. This is especially true with test servers because you’re constantly replacing
old component versions with new ones.
Always begin fixing this problem by registering your components and type libraries again,
and then rebooting the server. The server will often read a copy of your component into
memory, and fail to release it, even when you start and stop the Web server. Clearing the
server’s memory is one way to ensure you start with a fresh setup.

You can check for component registration and configuration errors by creating a local
test application and running it on the server. If the test application can successfully
interact with the component, the problem isn’t one of registration. In fact, you can
probably rule out component configuration problems as well.

Make sure you check for component configuration problems. For example, COM+ applica-
tions rely on a separate utility to configure the role-based security they use. Some compo-
nents require registry entries to determine where they are supposed to run. They normally
include a separate utility that you’ll use for configuration purposes.
Many SOAP toolkits require external configuration files that you need to check. For exam-
ple, 4S4C requires entries in a CONFIG.XML file that tell where to find the component
and which component to use. Unfortunately, if you make a mistake in this file, you need to
fix it and then stop and restart the Web server. Like the Microsoft SOAP Toolkit, 4S4C
reads its configuration files into memory to enhance performance.
Another persistent problem is one of Web server configuration. If you use virtual directo-
ries, make sure the directory setup is correct as well. For example, some operating systems
will prevent access to a virtual directory if you configure one set of security rules from the
Web server and another from the server’s file system configuration utility. Look for every
potential source of configuration error as part of your debugging strategy.
278 Chapter 8 Providing Remote Database Access

Addressing Transaction Issues

At the time of this writing, SOAP doesn’t provide a form of native transaction support.
Many developers feel this is a major concern because they’re used to using transactions for
every desktop application. The first question that you should ask yourself is whether your
application actually requires the use of transactions. Consider the importance of the data
and the worst-case scenario if the data is either lost or damaged. In at least some cases,
you’ll find that transactions aren’t required. For example, transactions aren’t necessary for
applications that only allow the user to view the data in question without modification.
Likewise, applications in which the user can only add or modify non-critical data can proba-
bly exist without benefit of transactions.
No matter what you do, many applications are still going to require transactions. The data
isn’t critical enough that you have to keep it off the Internet, but it’s also critical enough that
you need to protect it in some way. The big thing to remember is that you’ll incur data loss
when using SOAP because the Internet is inherently unsafe.

The Internet is unsafe. If you have any doubts about this statement, spend some time
reading about the security breaches that occur every day. If your data is so critical that
you can’t afford loss or damage, your best bet is to keep it off the Internet. This book
provides you with some ideas of how to circumvent problems with data transfer on the
Internet. However, none of these methods guarantees specific results, and you might
find that you still lose data. Always provide some means of recovering the data and
keep in mind that you’re going to lose some data along the way.

Some developers have come up with interesting ways to provide at least a modicum of trans-
action support for their applications. One technique that seems popular is to add three ele-
ments to every SOAP message: user ID, session ID, and package ID. Using these three
elements allows you to track the progress of every piece of data that your application trans-
fers. However, using this technique requires that you maintain state information outside of
the application. This means that there’s a chance the transaction information could get out
of sync with the actual data transfer and you’ll lose data anyway.
Another popular technique is to use cookies to track the progress of a data transfer.
However, this means that you must set up both systems to use the cookies and agree on a
methodology for using them. Some developers also complain that using the cookie method
is overly complicated.

Transactions are a hot issue for developers right now. You can find a wealth of information
on the topic online. TNL.NET (http://www.tnl.net/newsletter/2001/
soapsecurity.asp) includes a newsletter that addresses the issue of SOAP security
during transactions. You can find a preliminary demonstration of SOAP transactions at
http://www.xbrlsolutions.com/public/demos/crossreference/Soap/Post
 Troubleshooting 279

Transaction.asp. You’ll find at least one white paper on the topic of transactions on
the SOAP Web Resource Center (http://www.soap-wrc.com/webservices/
default.asp). Perl users will want to read the Quick Start with SOAP (http://www.
perl.com/pub/2001/01/soap.html) white paper that also includes information
about transaction support. A number of other sites offer small tidbits of information about
transactions for specific languages. Of course, you can always read about SOAP transac-
tions in the specification (http://www.w3.org/TR/SOAP/#_Toc478383497).

Obviously, these two techniques are inadequate for today’s data transfer needs. The vendors
involved with SOAP are working on a new set of specifications that will allow developers to
use transactions in a consistent fashion. After vendors standardize the method for using
transactions, you’ll be able to use SOAP for data that is more critical than your average task
list. At the time of this writing, transactions are more a gleam in the eye of some vendors
than something written down on paper. It will take some time for vendors to devise a good
solution to the problem of transactions—one that will honestly work for at least the majority
of SOAP users.

Troubleshooting
This chapter has shown you how to work with database applications under SOAP. It’s
important to understand that SOAP fixes some problems but presents other challenges.
Some developers consider these challenges so severe that they won’t use SOAP for database-
oriented tasks. SOAP does increase certain database-related risks as we’ve discussed.
However, those risks are small enough that you can use SOAP for some types of database
applications. The following sections will examine some of the questions developers ask
about using SOAP for database management and hopefully provide the answers you need to
add SOAP to your database management application toolkit. (Always feel free to contact me
at [email protected] if you have additional questions.)

How Do I Detect SOAP-Related Database Errors?

You should begin by using the same techniques that desktop applications use. Make sure you
create components that filter data and look for obvious errors, such as data that’s out of
range or doesn’t follow some other criterion. Transaction processing is the safest way to
transfer data from one point to another, but many developers use alternatives because trans-
actions exact a heavy performance toll. Until SOAP provides transaction support, your best
bet for finding problems is to perform detailed data analysis.
Of course, you have to have data to analyze, and SOAP isn’t guaranteed to get the data from
one point to another. In fact, the Internet is a decidedly unsafe way to transfer data because
the data can become lost at any point. Another method you want to use for database moni-
toring is to keep a local copy of the data on the user’s machine until the user receives verifi-
cation that the data ended up at the server in good condition. Most of the applications in
this book show how to use a response mechanism, so sending a receipt isn’t a problem.
280 Chapter 8 Providing Remote Database Access

You do need to consider a subtle problem when using receipts. The data could end up at the
server in good shape, but you might not receive the receipt. Perhaps the connection became
bad after the data transfer, but before the receipt was transmitted. Maintaining copies of
records on the machine is a good idea, but you shouldn’t transmit them again automatically.
Instead, mark these records for a manual check after the user gets back to the office to
ensure that the database contains only one copy of the data.
The best strategy for working with databases is to keep critical data off the Internet. If the
data is so important that there’s a 0% tolerance for losing it, then you need to use methods
that are safe to transfer it. This means using a local connection and a binary transfer proto-
col. Using SOAP to transfer critical data will result in an unrecoverable loss at some point.

Are the Database Shortcomings of SOAP Permanent?

Vendors are constantly working on improving SOAP. For example, you can read about even-
tual changes to how the SOAPAction element works at http://lists.w3.org/Archives/
Public/xml-dist-app/2001May/0026.html. Remember that this protocol is a standard in the
rough—it’s barely ready for use by the general public. Given the number of companies who
have already developed products for SOAP and the number of new products that will appear
on the scene soon, the future of SOAP looks bright. There’s little doubt that vendors will fix
the problems that SOAP has today and you’ll be able to use it for other purposes.
The two big issues that SOAP innovators need to solve to make SOAP a truly useful data-
base platform are transactions and security. SOAP will never transfer data with the 100%
assurance that companies require until it supports transactions. Likewise, given the open
nature of the Internet, vendors won’t use SOAP for critical data until it includes built-in
security. This means some form of encryption and digital certificate support. IBM has
already come out with one solution for the security problems at the time of this writing, but
it isn’t a standard solution. SOAP will need a standardized security solution to survive.
Some developers would add a third issue to this discussion. The SOAP specification cur-
rently promises support for attachments, but no one has implemented this support so far.
Several vendors are working on attachment support for their products, and you might see
some of those solutions on the market by the time you read this. It’s important that they
implement attachments using a standardized methodology, so it might be a while before you
see SOAP attachments that truly work across all platforms.

What Are the Top Ten Issues for SOAP Database Developers?
SOAP database developers need to exercise more care than any other SOAP user. SOAP is a
great protocol, but it’s hardly a tested product. The following list presents the ten issues that
I feel database developers need to consider the most when creating SOAP applications:
1. Always verify that you can convert database data to something with which XML will
work.
2. Always verify with which platforms your application will need to work because each
platform presents special challenges.
 Troubleshooting 281

3. Always use small test cases to test the viability of your database project before you com-
mit to the larger project.
4. Always validate the feature set of the SOAP toolkit because there’s a lot of hype that
doesn’t appear as usable functionality.
5. Always buffer the data to keep crackers at bay and keep the data safe.
6. Never allow third-party vendors direct access to your backend processing—use secure
intermediate components instead.
7. Never erase the local copy of the data before you verify that it has arrived on the server
in good condition.
8. Never allow users to remove data from a remote location—always move the data to a
holding database instead.
9. Reduce the effects of connection failure by maintaining an alternative connection
strategy, such as DCOM or CORBA.
10. Increase your chances of a successful project by using a high level of modularity of
project components.
 CHAPTER
9
Moving to Web-Based
Applications
In this chapter
Uses for Web-Based Applications 285
Overcoming Problems with Web-Based Applications 287
Updating a Thick Client Application for Thin Client Use 289
Creating a Live Data Application 302
Handling Web-Based Application Errors 303
Security Issues for Web-Based Applications 307
Quick Fixes for Memory and Other Resource Problems 312
Case Study 317
284 Chapter 9 Moving to Web-Based Applications

It wasn’t very long ago that everyone worked on a PC, alone in a world of their own. LANs
connected the PCs together, but people were still essentially working alone. Today, people
no longer work alone and it’s quite possible that they won’t use a PC to communicate.
Collaborations are becoming the rule of the day, and you’ll find that they occur on every-
thing from PCs to personal digital assistants (PDAs). In the near future, you might even use
your cell phone to perform certain computer-related tasks. Tying everything together is the
Internet and the Web-based application.
This chapter acquaints you with the Web-based application. In the first section of this chapter,
“Uses for Web-Based Applications,” we talk about how you can put these new applications
to work in your organization. You might not have a requirement for PDA communication
today because it’s a new technology, and only a few companies are brave enough to test new
technology waters before they’re completely proven. You’ll likely need to create Web-based
applications in the future, and this section helps you see their potential.
The second section of this chapter, “Overcoming Problems with Web-Based Applications,”
helps you understand how Web-based applications currently fall short of the ideal. What am
I using for comparison? The desktop application is the ideal by which users normally judge
computer software, so that’s what we’ll use in this section. Although you can overcome some
of the problems using clever programming, other problems will steadfastly refuse any form of
fix using current technology. This section also talks about those problems that you can
repair, and those that you’ll have to live with for now.
The third section, “Updating a Thick Client Application for Thin Client Use,” looks at the
problem most of you’ll have. You currently have a lot of desktop code that works fine on a
PC, marginally on a laptop, and not at all on a PDA. Developers commonly refer to desktop
applications as thick clients. Web-based applications use the thin client approach because it
places less of a burden on the client computer. This section looks at techniques you can use
to move a thick client application into the world of the thin client. We begin with a thick
client application and move it to a thin client environment. (We won’t look at a PDA example
in this section—that example appears in Chapter 10, “Working with PDAs.”)
Many applications today rely on live data, rather than static data. Unfortunately, Web pages
are too often associated with static data. It’s hard to keep data live using a Web connection
that might not work from one minute to the next. The “Creating a Live Data Application”
section of this chapter looks at techniques that you can use to create a live data application—
one where the data changes on the client as it changes on the server. It’s not always neces-
sary to provide live data, so we also talk about the situations where live data is most
appropriate. In short, the example application in this section is an alternative to working
with static data.
The fifth section of this chapter, “Handling Web-Based Application Errors,” shows you how
to deal with problem connections and component errors. This section looks at problems
that you’ll most likely encounter when working with Web-based applications. For example,
losing the connection to the server is a very real possibility in this situation. We examine
ways to detect the connection loss and attempt to reestablish it automatically.
 Uses for Web-Based Applications 285

For the purposes of this book, the term cracker will always refer to an individual that’s
breaking into a system on an unauthorized basis. This includes any form of illegal
activity on the system. On the other hand, a hacker will refer to someone who
performs low-level system activities, including testing system security. In some cases,
you need to employ the services of a good hacker to test the security measures you
have in place, or suffer the consequences of a break-in. This book will use the term
hacker to refer to someone who performs these legal forms of service.

The sixth section of this chapter, “Security Issues for Web-Based Applications,” deals with a Ch
major problem today. Any time you open a server to the Internet, you risk exposure to crackers.
However, placing a client on the Internet also poses subtle risks that many companies fail to 9
take seriously. This section looks at the security risks at both ends of the connection and
provides you with code to help secure your applications.
The seventh section of this chapter, “Quick Fixes for Memory and Other Resource
Problems,” looks at another serious Web-based application problem. Most of the client
devices for Web-based applications are memory or resource restricted. Even a laptop has
fewer resources, in most cases, than a desktop machine does. This section helps you under-
stand how to circumvent the limitations of Web-based application clients.
The final section of this chapter contains a case study showing how another company is
using Web-based applications to get work done today. Most companies are still in the exper-
imental stage with their Web-based applications as I write this. Eventually, these new appli-
cations will find their way into the mainstream and make it possible to interact with your
company using a larger variety of devices than ever before.

Uses for Web-Based Applications

For some people, moving to Web-based applications is a given because they have a large
sales staff on the road or some other remote communication concern. However, for many
people, the question of using Web-based applications might not have even come up yet.
Some developers find themselves dealing with the problem of the moment, rather than
looking at future requirements. The point is that Web-based applications are becoming
available, but they aren’t exactly common today. As technologies like SOAP become more
prevalent and stable, however, businesses will begin the move to the Internet in a big way.
Deciding which business processes to implement as Web-based applications isn’t easy.
Businesses have many concerns when it comes to moving data on the Internet, especially
after news reports of cracker activity make it apparent that this is a risky move. Of course,
moving anything in a company requires planning, but more importantly, you have to have a
goal or a vision for the project. Otherwise, it’s pointless to begin the process. The goal is to
find out which processes will work best on the Internet today, and wait for technology to
provide answers needed for other application types.
Some applications definitely work better than others do when modified for the Web-based
application environments. Eventually you’ll be able to move any application to the Internet,
286 Chapter 9 Moving to Web-Based Applications

but today, current technology limits what you can do. Social concerns also come into play.
The employees who will use these new applications will have to develop a new set of skills
and methods of looking at data. Web-based applications typically rely on a browser interface
that’s not equipped to provide a desktop application appearance.
One of the easiest applications to move to the Web-based application environment is the help
desk application. You can use help desk applications for more than simple or even complex
help files. Some companies use help desk-like applications to provide alternative information
to the user, not just help. For this reason, the help desk application is both more complex and
more flexible than the help file. The following list provides you with some ideas on how you
can use help desk applications to do more than just provide helpful information.
■ Company Policies: The policies used to run your company are normally stored in
printed format, making them difficult to access. Using a help desk application to make
this information available online is one way to reduce the time required searching for a
company policy, making it more likely that employees will follow such policies.
■ Bulletin Board: Companies usually have bulletin boards containing announcements,
employee regulations, and the like. In many cases, someone has to go from bulletin board
to bulletin board making updates as required. A single help desk application can replace
all of the company bulletin boards. Scanned images can replace paper counterparts. In
addition to saving time, using an electronic bulletin board allows more freedom in the
presentation of information (use of animation is just one example) and closer monitoring
of bulletin board content.
■ Forms: Finding a required form can be a difficult and time-consuming process. In fact,
this problem results in duplicate forms for many companies because each person is
certain that the form doesn’t exist. Using a help desk application to allow employees
to find forms can save time, money, and duplications.
■ Company Locator: Someone with a company telephone book has to know someone about
the company before he can use it. A help desk application that provides the same function-
ality, on the other hand, is useful from the very first moment. Instead of knowing that
James Smith is the personnel director, a user can simply ask for the personnel director and
the application will present the required information.

Of course, you can use Web-based applications for more than just help desk support. Web-
based applications fulfill a variety of other roles. For example, you can use them for the
same utility applications that we discussed in the “An Overview of Remote Access Utilities”
section of Chapter 6, “Creating Remote Access Utilities.”
You’ll need to provide additional security for the Web-based versions of utility programs
because it’s likely that you’ll always use them from a remote location. The Web-based version
of these applications will also look and act slightly different. This could actually become an
advantage because you could take advantage of technologies such as eXtensible Hypertext
Markup Language (XHTML) to allow the application to run on multiple hardware platforms.
Imagine the benefits of administering a network from a cellular telephone.
 Overcoming Problems with Web-Based Applications 287

Some companies are creating full-fledged applications using a Web-based format. For exam-
ple, some developers feel that new versions of word processors will rely on the Web-based
format instead of using proprietary document formats as they do now. (See the “Determining
Which Data Entry Vehicle to Use” section of Chapter 7, “Creating Data Entry Forms and
Surveys,” for a discussion of StarOffice—one of the potential candidates for this transition.)
Using XML as a basis for storing all data means that you no longer have to worry about data
becoming inaccessible as the applications that created the data become older. Unfortunately,
using XML for all data storage also means investing in larger hard drives.

Ch
Many developers will become involved in embedded systems as embedded systems start
to appear in more places. You’ll currently find embedded systems in your home, car, and
workplace. Some vendors have already connected embedded systems such as alarms,
9
temperature monitoring, and car locators to remote networks, including the Internet. Java
and other high-level programming languages are becoming more popular for embedded
systems, making it easier for more companies to get involved. Added to this mix is a new
SOAP implementation for embedded systems from eSOAP (http://www.embedding.
net/eSOAP/). You’ll find an article about this product and its use at http://www.
embedding.net/eSOAP/english/index_script.html?src=%22Documents%22.
Your next project might not be on the PC—it could be someone’s smart toaster.

Finally, Web-based applications will eventually fill a new class of application. Information
exchange has become a central focus for application developers today. Look for Web-based
applications to extend this idea in the future. For example, it’s already possible to translate
human speech from one language to another. A Web-based application that made that feature
available on a PDA or telephone would be extremely helpful for business travelers. Likewise,
Web-based applications will eventually make it possible to perform client research onsite. A
salesperson could perform a credit check on a customer without ever leaving the area.

Overcoming Problems with Web-Based

Applications
Web-based applications present more challenges than the typical application because there are
so many variables to consider. For example, companies now demand that Web-based applica-
tions work on more than one hardware platform. You might find yourself testing an application
on both a desktop machine and a personal digital assistant (PDA). The differences in screen
size, capability, and operating system will make such testing interesting to say the least.
Browser compatibility is another issue that you’ll have to learn to circumvent. Web-based
application development is hard enough when everyone uses the same browser on the same
platform. The fact is that although Microsoft does enjoy a commanding lead in the browser
market, it doesn’t own the entire market. In addition, you’ll find that the different versions
of Internet Explorer are incompatible with each other. Even if everyone decides to use the
same browser, the chances of getting everyone’s version the same are slim.
288 Chapter 9 Moving to Web-Based Applications

The usual host of problems is present. In fact, you’ll probably find them amplified in the
Web-based application arena. If you have performance problems when working with appli-
cations on the desktop, the problems will be far worse when working with a Web-based
application. Lost connections and users who fail to follow procedures won’t help matters.
You can overcome most of these obstacles given time and a few resources. Unfortunately,
most developers are short on both time and resources. You cannot afford the time to conduct
extensive tests, hold users’ hands, and generally clear the environment of hazards. With this
in mind, the following list provides some quick problem fixes you should try during your next
Web-based application development session:
■ Decide in advance which platforms you’ll support. Make sure you consider the platforms
used by partner organizations. You’ll also want to define specifications that match current
company plans for upgrades and the installed base of products. It’s important to publish
a list of new equipment requirements as part of your specifications so you don’t continue
to struggle to support outdated products.
■ Test your application on all of the hardware platforms you expect to use. Performing
representative testing, where you check just a subset of the platforms, won’t work. The
number of SOAP toolkits on the market right now that work with just one PDA, work-
station, or server prevents representative testing from working.
■ Set up a hands-on lab for users. This will allow users to gain experience with the new sys-
tem before it suddenly appears on their desktop or other hardware. Make sure you include
several pieces of hardware in the lab so users can see differences between desktop and PDA
use. Stress that none of the data the user creates will remain after the system goes online.
This helps prevent some users from monopolizing the machines to get a good start on the
new product in advance. A lab also allows you to gather usability information from users
before the system goes online so that you can make any required changes early in the
design phase.
■ Test all connection types. More than a few companies have created a new application
and tested on their LAN. The application works great when used locally, but runs too
slow from a dial-up connection. Even if you don’t plan to use dial-up connections for
your application, try using your application from a dial-up connection anyway. It’s
important to consider all of your options when working with Web-based applications.
■ Try different coding techniques early in the design and implementation phases. SOAP
toolkits normally provide several ways of getting any job done. Choosing the most con-
venient method won’t always ensure good results. You might find that when you use the
Microsoft SOAP Toolkit, for example, you need to use the low-level API to gain a
performance advantage for your application.
■ Mix communication techniques as needed. SOAP isn’t an all-or-nothing solution to
coding problems. Many applications will require continued use of binary protocols,
 Updating a Thick Client Application for Thin Client Use 289

such as DCOM and CORBA, to provide adequate performance. This is especially true
for Web-based applications where performance is critical.
■ Diagram your application, components, and even message formats. An optimal setup
includes defining all application elements at the outset of the project, and then sticking
with the plan unless there’s a good reason to make a change. We’ll explore the BizTalk
business-to-business (B2B) solutions in Appendix B, “Microsoft BizTalk and SOAP.” A
major part of BizTalk is the utilities it provides for designing an application.
■ Consider adding automation to your application. Sometimes taking the human out of
the picture can produce large increases in efficiency and keep users happy as well. The
user will want to concentrate on his job, not on your application. Making tasks easy for Ch
the user usually garners positive results in application usage.
9
Updating a Thick Client Application for Thin
Client Use
Web-based application development revolves around the thin client. A thin client is one that
relies on the server to perform most of the processing. The thin client displays data and col-
lects information from the user, but doesn’t process the information. Thin clients are akin to
the dumb terminals used with mainframes in days gone by. Thin clients ddo possess more
intelligence, but they use the intelligence to service user needs rather than process data.
Contrast this with the thick clients that many developers are used to creating. A thick client
processes much information locally and passes only the results to the server. Instead of working
with raw data, the server merely acts as a storage and coordination device that tracks the state
of the completed information. Thick clients work well on peer-to-peer networks because each
machine contributes toward the overall processing goals. However, thick clients also require
many local resources, something that small devices such as PDAs and cellular telephones don’t
provide. The main reason that developers are moving away from thick clients in some situations
is that users are starting to use lower power devices to run applications.
On one end of the question, you have thick clients and on the other, you have thin clients.
However, the question isn’t black or white. Many developers see clients as a continuum from
thick to thin. Some clients exist in the middle by performing some processing locally and
asking the server to do the rest. The point of updating a client from thick to thin is to change
the location of the processing. The same processing takes place, but the location has changed.
The following sections will take you through the process of changing a thick client to a thin
client. You’ll also change the viewing application from a desktop application to a browser.
The same processing will take place after the change, but the way the application displays
the data and the location of the data manipulation will change.
290 Chapter 9 Moving to Web-Based Applications

Many developers are used to creating components with properties. Using properties
allows a developer to set up a component before executing commands. In this way, the
developer can perform a setup once and then make calls without including a long list of
arguments for each call. Unfortunately, properties won’t work with SOAP because SOAP
is stateless. However, you can use the two-component approach explained in this sec-
tion of the chapter to allow property usage within your SOAP application. A processing
component that always resides on the server sets the properties within the server com-
ponent using standard COM calls. A cookie identifies individual users so the processing
component can keep the calls separate. Using this technique allows you to retain some
of the benefits of using properties while working with SOAP applications.

Generating Proxy Classes the Easy Way

Third-party developers are constantly looking for ways to decrease the complexity of working
with SOAP. One technique is to create code generators that solve at least simple problems. You
can download the VBWS Proxy Generator that creates a Visual Basic class based on the con-
tent of your WSDL file at DevXpert (http://www.devxpert.com/resources/). This site con-
tains a wealth of other tools and helpful tutorials as well.
The VBWS Proxy Generator consists of three components: a configuration file that lists all
of the data types used in your application, an HTML application that provides a GUI, and
an XSLT file that performs the conversion. You start the VBWS Proxy Generator by
double-clicking the VBWS.HTA file. Figure 9.1 shows what you’ll see.

Figure 9.1
The VBWS Proxy
Generator features an
easy-to-use interface.
 Updating a Thick Client Application for Thin Client Use 291

All you need to do is point the utility to the WSDL file you created for your component
using the WSDL generator and supply a directory to hold the resulting Visual Basic class
file. Click Generate VB Code and the program will output a class file similar to the one
shown here. (This example shows the output from the AddIt component found in Chapter
4, “Using SOAP to Create a Simple Application.”)
Option Explicit

Private Const WSDL_URL As String =

“http://winserver/soapexamples/SimpleAdd/AddIt.wsdl”

Ch
Public Function DoAdd(ByRef Add1 As Integer, ByRef Add2 As Integer) As Integer
Dim soap As soapClient
Set soap = New soapClient
9
soap.mssoapinit WSDL_URL, “AddIt”, “AddItClassSoapPort”
DoAdd = soap.DoAdd(Add1, Add2)
End Function

All you need to do is add this class file to a component, and then use the component to
access the functionality of the server-side component. We looked at a similar example in the
“Updating a Simple Utility Program” section of Chapter 5, “Migrating an Application from
DCOM to SOAP.” The VBWS Proxy Generator automates this process to a certain extent.
Note that you still need to add error handling. In addition, the VBWS Proxy Generator
might not work for complex data-handling needs. You also need to work with the configura-
tion file to allow it to handle complex data types.

Using Thick and Thin Clients Simultaneously

Some organizations will require more than one client for their applications. Users will want
a desktop client (thick) for local use and a Web-based client (thin) for remote use. You can
create both clients with little additional work by remembering that the difference between
thick and thin is where the processing occurs.
Component technology is wonderful because you can break an application into as many
pieces as you need. Creating a processing component, the embodiment of the difference
between thick and thin is one way to handle the problem of multiple client types. You can
simply place the processing component on the user’s machine when creating a thick client.
Likewise, the server will hold the processing component when working with a thin client.
This still leaves a problem with communication. Remote clients will want an application
that communicates reliably over the Internet. To do this, you’ll need to use a protocol such
as SOAP for at least the thin client’s connection to the server. Another way to look at this
problem is convenience. If you design the processing component to use SOAP as both the
input and output protocol, you can move it anywhere and still communicate with the server.
That’s just what we’ll do in this example. The thick client is nothing more than a shell that
calls on a local processing component using SOAP. The thin client is nothing more than a
shell that calls on the remote processing component using SOAP. Except for the differences
in presentation, you’ll find that the applications operate about the same.
292 Chapter 9 Moving to Web-Based Applications

Creating the Server-Side Component

The server-side component for this example isn’t that complicated. All it does is detect the
name of the remote server and return it in a string. The server-side component could just
as easily access a database, obtain server status information, or perform other tasks as other
components in the book have done. You can use this component to obtain several common
computer name values available with the GetComputerNameEx() API function, as shown in
Table 9.1.

Accessibility to the GetComputerNameEx() API function didn’t ship with Visual Studio 6,
so you won’t find it in the copy of Microsoft Developer Network (MSDN) that ships
with that product. GetComputerNameEx() function access does come with the latest
version of the Platform SDK and should come with Visual Studio 7. The source code in
the \Chapter 09\Computer Name Component directory in the source code file available
from the Que Web site contains a module with the necessary function declaration. You
can find the source code at http://www.quepublishing.com. All other resources
appear with the component source code in this section. In short, even if you don’t find
the function in your current copy of MSDN, you’ll find everything needed to use it in
this component.

Table 9.1 Name Value Return Types for GetComputerNameEx()

Name Type Constant Description

ComputerNameNetBIOS Returns the NETBIOS name of the computer. If the
computer is part of a cluster, GetComputerNameEx()
returns the name of the cluster, rather than the
name of the individual computer.
ComputerNameDnsHostname Returns the DNS host name of the computer. If the
computer is part of a cluster, GetComputerNameEx()
returns the name of the cluster, rather than the
name of the individual computer.
ComputerNameDnsDomain Returns the DNS domain name assigned to the
computer. If the computer is part of a cluster,
GetComputerNameEx() returns the name of the
cluster, rather than the name of the individual
computer.
ComputerNameDnsFullyQualified Returns the fully qualified DNS name that uniquely
identifies the computer within the network as a
whole. This normally includes the computer name
and the name of the domain to which it belongs as
<HostName.DomainName>. If the computer is part
of a cluster, GetComputerNameEx() returns the
name of the cluster, rather than the name of the
individual computer.
ComputerNamePhysicalNetBIOS Returns the NETBIOS name of the computer,
even if the computer belongs to a cluster.
 Updating a Thick Client Application for Thin Client Use 293

Table 9.1 Continued

Name Type Constant Description
ComputerNamePhysicalDnsHostname Returns the DNS host name of the computer, even
if the computer belongs to a cluster.
ComputerNamePhysicalDnsDomain Returns the DNS domain name assigned to the
computer, even if the computer belongs to a cluster.
ComputerNamePhysicalDnsFullyQualified Returns the fully qualified DNS name that uniquely
identifies the computer within the network work as
a whole. It returns the name of the local computer
even if the computer belongs to a cluster. Ch

9
The enumeration also contains a ComputerNameMax value. This value is included
with the Visual C++ header file, and Microsoft documents it as part of the
COMPUTER_NAME_FORMAT description in the platform SDK. However, this value
is currently unused and will return a blank value.

The GetComputerName() function requires three arguments: one of the name type constants
shown in Table 9.1, a buffer used to hold the return value, and the length of the buffer. The
buffer length variable returns with the actual length of the name on return from the func-
tion call. GetComputerName() returns a Boolean value that’s true if the call is successful.
Now that we’ve discussed the basics, let’s look at the component code. Listing 9.1 shows the
component-specific code for this example. It doesn’t show the function declaration in the
GetCompNameMod.bas file.

Listing 9.1 CompName Component Source Code

‘An enumeration of the possible name types.
Public Enum COMPUTER_NAME_FORMAT
ComputerNameNetBIOS
ComputerNameDnsHostname
ComputerNameDnsDomain
ComputerNameDnsFullyQualified
ComputerNamePhysicalNetBIOS
ComputerNamePhysicalDnsHostname
ComputerNamePhysicalDnsDomain
ComputerNamePhysicalDnsFullyQualified
ComputerNameMax
End Enum

Public Function GetCompName(NameType As COMPUTER_NAME_FORMAT) As String

‘Create and initialize a computer name buffer.
Dim NameBuffer As String
NameBuffer = Space(80)
294 Chapter 9 Moving to Web-Based Applications

Listing 9.1 Continued

‘Create a buffer length variable.
Dim BufferLength As Integer
BufferLength = 80

‘Get the computer name based on input values.

If GetComputerNameEx(NameType, NameBuffer, BufferLength) Then

‘Check the Buffer Length

If BufferLength > 0 Then

‘Return the computer name to the caller.

GetCompName = NameBuffer

Else

‘Return a no name value.

GetCompName = “No Name Assigned”

End If

Else

‘Return a generic name.

GetCompName = “No Name Available”
End If

End Function

Public Function GetAllNames() As Variant

‘Create an array to hold the values.
Dim AllNames() As String
ReDim AllNames(7, 1)

‘Fill the array with values.

AllNames(0, 0) = “ComputerNameDnsDomain”
AllNames(0, 1) = GetCompName(ComputerNameDnsDomain)
AllNames(1, 0) = “ComputerNameDnsFullyQualified”
AllNames(1, 1) = GetCompName(ComputerNameDnsFullyQualified)
AllNames(2, 0) = “ComputerNameDnsHostname”
AllNames(2, 1) = GetCompName(ComputerNameDnsHostname)
AllNames(3, 0) = “ComputerNameNetBIOS”
AllNames(3, 1) = GetCompName(ComputerNameNetBIOS)
AllNames(4, 0) = “ComputerNamePhysicalDnsDomain”
AllNames(4, 1) = GetCompName(ComputerNamePhysicalDnsDomain)
AllNames(5, 0) = “ComputerNamePhysicalDnsFullyQualified”
AllNames(5, 1) = GetCompName(ComputerNamePhysicalDnsFullyQualified)
AllNames(6, 0) = “ComputerNamePhysicalDnsHostname”
AllNames(6, 1) = GetCompName(ComputerNamePhysicalDnsHostname)
AllNames(7, 0) = “ComputerNamePhysicalNetBIOS”
AllNames(7, 1) = GetCompName(ComputerNamePhysicalNetBIOS)

‘Return the result to the client.

GetAllNames = AllNames
End Function
 Updating a Thick Client Application for Thin Client Use 295

As you can see, the component contains an enumeration for the computer name type. The
enumeration makes it easier to remember which computer name type values to use.
The two functions are relatively simple. The GetCompName() function does most of the work.
Remember to initialize the buffer before using it with an API call. The GetCompName() func-
tion also checks for two potential error conditions and provides alternative values. The
GetAllNames() function simply creates an array of values by calling GetCompName() multiple
times. This reduces the number of calls that a remote client needs to make to get all of the
name information for a server.
You should notice two problems with this component—problems that are common with many
Ch
components on the market. The GetCompName() function requires an enumerated type as input,
while the GetAllNames() function uses a variant as output. Neither of these functions will work 9
as written for a SOAP call because most SOAP toolkits can’t handle complex types of this
sort. The processing component (see the next section) will handle translating these function
calls into something that is easier to use. You should consider the processing component as
the component that performs translation as well as provides a means to move around the
processing needs of a component.

Creating the Processing Component

The processing component for this example has several problems to overcome. First, it
needs to overcome the problem of data typing. One way to do that is to use a third-party
alternative as we did for the complex data type example in Chapter 8, “Providing Remote
Database Access.” (See the “Using Complex Data Types” section of chapter 8 for details.)
Another method is to convert the data type into something compatible with SOAP. Finally,
we can perform a complete rewrite of the data into data the client has to parse later.
All of these methods are good. They work within a limited range and you’ll find that they’re
successful most of the time. The big problem is that these techniques reduce the readability
of your application. You no longer have a verifiable connection to the server component
because the means for calling it has changed.
The component in this section not only shows server-side processing techniques, but it also
shows two new techniques you can use to overcome the problems with data type conversion. Of
course, these techniques are just new tools to add to your existing toolkit—they won’t replace
the methods you already use. Listing 9.2 shows the processing component source code.

Listing 9.2 CompNameProc Source Code

Public Function GetCompName(ByVal NameType As Integer) As String
‘Create an instance of the CompName component.
Dim MyCompName As CompName.NameValues
Set MyCompName = New CompName.NameValues

‘Create a conversion variable for the incoming integer data.

Dim ConvNameType As COMPUTER_NAME_FORMAT
ConvNameType = NameType
296 Chapter 9 Moving to Web-Based Applications

Listing 9.2 Continued

‘Return the computer name data to the client.
GetCompName = MyCompName.GetCompName(ConvNameType)
End Function

Public Function GetAllNames() As String

‘Create an instance of the CompName component.
Dim MyCompName As CompName.NameValues
Set MyCompName = New CompName.NameValues

Dim Result As Variant ‘Contains an array of result strings.

Dim Counter As Integer ‘Loop processing variable.
Dim ParseString As String ‘Output string for client.

‘Get an array of computer name values.

Result = MyCompName.GetAllNames

‘Create a heading for the result string.

ParseString = “Computer Name” & Space(27) & “Value” & vbCrLf & vbCrLf

‘Process the array. Make sure you account for null-terminated string
‘values as part of the processing.
For Counter = 0 To 7
If InStr(Result(Counter, 1), Chr(0)) > 0 Then
ParseString = ParseString & Result(Counter, 0) & _
Space(40 - Len(Result(Counter, 0))) & _
Left(Result(Counter, 1), &_
InStr(Result(Counter, 1), Chr(0)) - 1) & _
vbCrLf
Else
ParseString = ParseString & Result(Counter, 0) & _
Space(40 - Len(Result(Counter, 0))) & _
Result(Counter, 1) & vbCrLf
End If
Next

‘Return the formatted data string to the client.

GetAllNames = ParseString
End Function

The first function, GetCompName(), accepts an integer as input. The integer takes the place
of the COMPUTER_NAME_FORMAT enumeration used by the server-side component.
However, we still need to convert the integer into a COMPUTER_NAME_FORMAT vari-
able the server-side component will understand. That’s easily accomplished, in this case, by
assigning the integer input to a COMPUTER_NAME_FORMAT variable. We can’t send
this value directly to the server without generating an error—you must perform the conver-
sion separately. After the code performs the conversion, it calls GetCompName() as normal.
The GetAllNames() function begins by creating a connection to the server-side component
and obtaining the array containing the return values. We’ll need to parse the data into a
string. The parsing process will also format the data. This differs from other components in
the book. The purpose of the processing component is to provide the data to the client
 Updating a Thick Client Application for Thin Client Use 297

application in ready-to-display format. You can assume that creating specific output is required
in this case. The idea is to reduce the size of the client and the amount of processing it
must perform.
At this point, we need to take a close look at the data returned by the server-side component.
You’ll encounter an odd problem when working with API calls. Microsoft wrote the API calls
in C/C++, which means they use null-terminated strings. Notice the bit of code required in
the two functions to overcome this problem. You need to check for the null character and
eliminate it before you can pass it along for the client. This isn’t as big of a problem for
GetCompName() as it is for GetAllNames(), but both functions could experience problems.
Ch
Designing the Thick Client Application
Designing the thick client is going to be relatively easy because the processing component 9
performs nearly all of the work required to prepare the data for display. Listing 9.3 shows
the source code for the thick client.

Listing 9.3 Thick Client Source Code

Private Sub cmdGetAllNames_Click()
‘Create the processing object.
Dim ComputerName As CompNameProc.NameValuesProc
Set ComputerName = New CompNameProc.NameValuesProc

‘Get the data.

txtResult.Text = ComputerName.GetAllNames
End Sub

Private Sub cmdGetSingleName_Click()

‘Create the processing object.
Dim ComputerName As CompNameProc.NameValuesProc
Set ComputerName = New CompNameProc.NameValuesProc

‘Get the data.

txtResult.Text = ComputerName.GetCompName(comboName.ListIndex)
End Sub

Private Sub cmdQuit_Click()

‘Exit the application.
End
End Sub

Private Sub Form_Load()

‘Set the list index value.
comboName.ListIndex = 0
End Sub

As you can see, the client code consists of creating the processing object and displaying the
result. This two-step process is all that the client should have to do at this point. The pro-
cessing component should perform all required data formatting and handling. Figure 9.2
shows the output from this application.
298 Chapter 9 Moving to Web-Based Applications

Figure 9.2
The thick client appli-
cation accepts the
preprocessed data
from the processing
component.

In this case, processing still occurs on the client machine. You’ll register the processing com-
ponent on the client and the processing component will access the server-side component
using DCOM. In short, you’ve isolated the processing portion of the application.

Designing a Thin Client Form View

Before moving to a Web page, it’s a good idea to test thin client capabilities using a form view
client. This allows the developer to check the viability of the client using code he can debug
easily. Remote access applications present enough challenges that you don’t want to try
debugging Web page code while checking component availability. This middle step reduces
the amount of debugging you’ll need to perform and increases the chances of getting the
Web page right the first time. Listing 9.4 shows the code for this part of the example.

Listing 9.4 Form View Thin Client Source Code

Private Sub cmdGetAllNames_Click()
‘Create the SOAP client.
Dim Client As SoapClient

‘Set up an error handler.

On Error GoTo ErrorHandler

‘Create the connection.

Set Client = New SoapClient
Client.mssoapinit _
“http://winserver/soapexamples/ComputerName/CompNameProc.WSDL”, _
“CompNameProc”, _
“NameValuesProcSoapPort”

‘Get the data.

Dim ParseString As String
ParseString = Client.GetAllNames
ParseString = Replace(ParseString, Chr(10), vbCrLf)
txtResult.Text = ParseString

Exit Sub

‘Display a message when an error occurs.

ErrorHandler:
 Updating a Thick Client Application for Thin Client Use 299

Listing 9.4 Continued

MsgBox Client.faultstring, vbExclamation
End Sub

Private Sub cmdGetSingleName_Click()

‘Create the SOAP client.
Dim Client As SoapClient

‘Set up an error handler.

On Error GoTo ErrorHandler

‘Create the connection. Ch

Set Client = New SoapClient
Client.mssoapinit _
“http://winserver/soapexamples/ComputerName/CompNameProc.WSDL”, _
9
“CompNameProc”, _
“NameValuesProcSoapPort”

‘Get the data.

txtResult.Text = Client.GetCompName(comboName.ListIndex)

Exit Sub

‘Display a message when an error occurs.

ErrorHandler:

MsgBox Client.faultstring, vbExclamation

End Sub

Private Sub cmdQuit_Click()

‘Exit the application.
End
End Sub

Private Sub Form_Load()

‘Set the list index value.
comboName.ListIndex = 0
End Sub

As you can see, this code follows the same two-step process as the thick client. In this case,
we establish a connection with the server, then use the required function to display the data.
It’s important to remember that the XML parser will still strip the carriage returns from
your data stream despite being a valid character according to the specification. Using the
Replace() function allows you to restore the carriage return/linefeed combination without
too many problems.

Designing the Thin Client Web Page

Many of your SOAP clients will probably include Web page design at some level. This sec-
tion shows how to create a basic Web page client that relies on scripts to accomplish its work.
The Web page thin client uses a slightly different technique to create a SOAP client, but the
process is essentially the same. Listing 9.5 shows the source code for this part of the example.
300 Chapter 9 Moving to Web-Based Applications

Listing 9.5 Web Page Thin Client Source Code

<HTML>
<HEAD>
<TITLE>CompName VBScript Example</TITLE>

<SCRIPT LANGUAGE=”VBScript”>
Sub cmdGetSingleName_Click()
‘Create a form reference.
Dim Form1
Set Form1 = Document.SampleForm1

‘Create the SOAP client.

Dim Client
set Client = CreateObject(“MSSOAP.SoapClient”)

‘Get the data and display it.

Client.mssoapinit
“http://winserver/soapexamples/ComputerName/CompNameProc.WSDL”, _
“CompNameProc”, _
“NameValuesProcSoapPort”
Form1.Results.Value = Client.GetCompName(CInt(Form1.comboName.Value))
End Sub

Sub cmdGetAllNames_Click()
‘Create a form reference.
Dim Form1
Set Form1 = Document.SampleForm1

‘Create the SOAP client.

Dim Client
set Client = CreateObject(“MSSOAP.SoapClient”)

‘Get the data and display it.

Client.mssoapinit
“http://winserver/soapexamples/ComputerName/CompNameProc.WSDL”, _
“CompNameProc”, _
“NameValuesProcSoapPort”
Form1.Results.Value = Client.GetAllNames
End Sub
</SCRIPT>

</HEAD>

<BODY>
<!Add a heading->
<CENTER><H1>Computer Name Component Test</H1></CENTER>

<!Use a form to test out the VBScript.->

<FORM NAME=”SampleForm1”>

<!Create a text entry control.->

Select a single computer name if needed:<BR>

<select name=comboName>
 Updating a Thick Client Application for Thin Client Use 301

Listing 9.5 Continued

<option value=0>ComputerNameNetBIOS
<option value=1>ComputerNameDnsHostname
<option value=2>ComputerNameDnsDomain
<option value=3>ComputerNameDnsFullyQualified
<option value=4>ComputerNamePhysicalNetBIOS
<option value=5>ComputerNamePhysicalDnsHostname
<option value=6>ComputerNamePhysicalDnsDomain
<option value=7>ComputerNamePhysicalDnsFullyQualified
<option value=8>ComputerNameMax
</select><P><P>
Ch
<!Define a place to put the results.->
Result values: <BR>
<TEXTAREA ROWS=11 COLS=60 NAME=”Results”>
9
</TEXTAREA><P><P>

<!-Create the two required buttons.->

<INPUT TYPE=button
NAME=cmdGetSingleName
VALUE=”Get Single Name”
ONCLICK=cmdGetSingleName_Click()>
<INPUT TYPE=button
NAME=cmdGetAllNames
VALUE=”Get All Names”
ONCLICK=cmdGetAllNames_Click()>

</FORM>
</BODY>
</HTML>

As you can see, the HTML page uses standard tags to create the document. The two but-
tons have the same names and function as the other clients. Note that each button calls a
separate VBScript function that bears a remarkable resemblance to the code we’ve used for
other clients.
You have to create the client differently due to limitations in scripting code. After declaring
the Client variable, the code uses the CreateObject() function to instantiate a SoapClient
object. After you have an object to workwith, you’ll use mssoapinit() as normal to initialize
the object. Make sure you include the double quotes as part of the call.
Displaying the page means keeping security low because the components aren’t marked safe
for scripting. You’ll see a warning message similar to the one shown in Figure 9.3 when you
click either of the buttons.
Click Yes and the call will proceed. Figure 9.4 shows the final application output when click-
ing Get All Names. Notice that the client code didn’t perform formatting, yet the output
looks the same as the thick client code shown in Figure 9.2. This is one of the side benefits
of using a processing component. The processing component formats the output to all
clients the same, so each client maintains a consistent look.
302 Chapter 9 Moving to Web-Based Applications

Figure 9.3
The thin client appli-
cation will display a
warning message
because the compo-
nent isn’t marked safe
for scripting.

Figure 9.4
The output of the thin
client looks similar to
the thick client due to
the processing com-
ponent output.

Creating a Live Data Application

At least a few Web pages now tout dynamic content. In other words, the content of the Web
page changes to match real-world events. In most cases, these Web pages rely on advanced
programming techniques and client-side components to do their job. In all cases of true
dynamic (or live) content, the client and server require several features as listed below:
1. The server component must generate events based on a real-world occurrence. For
example, when the price of a stock changes, the server-side component monitoring that
price will generate an event.
2. The client application must provide some means for accepting the event. It must listen
for the server’s call.
3. The server must provide some means to detect live connections or store data for clients
that have gone offline. For example, the server could use a publish/subscribe model or
store the requested data in a server-side queue. The client can either opt out of the data
flow or request stored messages when contacts the server again.
 Handling Web-Based Application Errors 303

4. The server must provide some method to filter unwanted data. Sending every piece of
information to all clients will waste bandwidth and could cause memory-limited clients
such as PDAs to crash. Even when the client doesn’t crash, the user certainly won’t
want to wade through reams of useless information.

SOAP doesn’t provide any of these features. In addition, downloading the components to
create a dynamic Web page runs counter to the reason for using SOAP. You’ll find that
downloading a component using SOAP is time-consuming, resource intensive, and difficult.
Therefore, your chances of creating a Web page with true dynamic content are slim.
However, you can copy the techniques used by other Web authors to create content that
Ch
looks dynamic, but really isn’t. The following list provides some ideas that you could pursue:
■ Use local timers to force a polling sequence by the client. Polling is generally frowned
9
upon as resource intensive and for wasting bandwidth; however, polling at reasonable
intervals is acceptable for custom applications. (You can determine a reasonable polling
interval by considering the amount of data the client downloads, the processing load on
the server, and the amount of available bandwidth.)
■ Use push technology to force a client update. Of course, push technology is only effec-
tive if the client allows it; many companies set their browsers to maintain strict security.
Limitations in the way push technology works is one of the reasons that vendors have
looked for better solutions.
■ Allow client-initiated updates. You could add a button to the Web page that would
allow the client to download new content as needed. However, this is hardly automatic,
and the client might push the button often out of frustration when waiting for an
answer to a query.

Will SOAP eventually provide a means for creating dynamic Web pages? Nothing is in the
works right now, and few developers have even asked about it. The few who have asked
are in businesses that legitimately require dynamic content, such as the stock market. It’s
unlikely that you’ll see true dynamic content anytime soon, so these pseudo-dynamic
solutions are all that is available.

Handling Web-Based Application Errors

Web-based applications suffer all of the common SOAP problems that we’ve discussed so
far in this book. For example, a Web-based application is just as sensitive to changes in the
WSDL file as a desktop application is. You don’t get special functionality using a Web-based
application either. Nothing about a Web-based application will make it magically handle the
complex data type scenarios that we discussed in Chapter 8.
Developers face some additional challenges when working with Web-based applications. For
example, unlike your LAN, the Internet has connectivity problems. The following sections
discuss some potential application errors that you’ll need to overcome to make your Web-
based application work properly.
304 Chapter 9 Moving to Web-Based Applications

Handling Connection Loss

Loss of connection for a SOAP application isn’t the same as loss of connection for DCOM
or CORBA applications. Remember that SOAP applications are one-time communications;
the server doesn’t maintain state or any other information about the client after it responds
to a request. That’s why you have to maintain state information outside of the application.
It’s also the reason that a loss of connectivity isn’t a major concern. SOAP applications are
actually more robust than DCOM or CORBA applications in this regard.
Detecting loss of connection is relatively simple. The client application will receive a No
HTTP response received message, like the one shown in Figure 9.5. The client application
can detect such a message and automatically send the request again. The application should
include a retry count so that it doesn’t try to send the request indefinitely.

Figure 9.5
A No HTTP
response
received message
indicates loss of
connection in
most cases.

Applications can lose a connection for a number of reasons. In some cases, you might want
to go further than simply sending the request again. For example, when you set up an appli-
cation for use while an employee is on the road, you might want to add code that checks for
a dial-up connection. The client could have lost the connection due to a problem with the
modem or other hardware. The point is to check beyond the application as needed when
using a connection that isn’t of the highest quality.

Scripting Errors
It’s likely that your Web-based application will rely on a browser—it won’t be a self-contained
executable like a desktop application. This means you’ll use scripts to perform tasks, such as
creating the connection and making a request.
The problem with scripts is that they’re unreliable, even when fully debugged and full of
error-trapping code. Think for a moment about any number of Web sites that you’ve
visited. I have often fixed scripting errors on these pages by clicking Refresh and waiting
for the page to reload.
As a developer, this means you need to spend more time debugging your scripts and ensuring
that a simple click of the Refresh button does the trick. Part of your employee training should
show how to refresh a page when needed.
You also need to exercise care in the selection and upkeep of browsers for your application.
Using a single browser for the entire company is a good start. However, you’ll want to go fur-
ther and ensure everyone has the same version of the browser with the same patches installed.
 Handling Web-Based Application Errors 305

Many developers underestimate the number of browser-related problems they’ll have with
an application. This truly is the weakest link of a Web-based application, and you need to do
everything possible to overcome it.

Component Security Problems

You’ll more than likely spend a lot of time handling security problems with a Web-based
SOAP application. Depending on the toolkit that you use, the vendor might not have
marked the components safe for scripting. This means that you’ll have to wrap the SOAP
component in a script safe component, use the SOAP component as a lower security level,
or write your own version of the client portion of the SOAP component. Of the three Ch
options, marking the component safe for scripting (assuming it’s safe for scripting) is the
easiest and most efficient route to take. 9
If you look at the examples in the Microsoft SOAP Toolkit, you’ll see another method of
overcoming the scripting problem. You create multiple files, some of which reside on the
server. The client never sees anything but HTML. The SOAP code executes on the server,
which would seem to defeat the purpose of using SOAP in the first place. This method does
work and you’ll find that you don’t need to mark components safe for scripting to use it.
However, the other methods mentioned in the section perform better. They also make
better use of SOAP and allow you to control the flow of information.
Whether you use SOAP or straight HTML for your application, the data you transfer
between client and server is at risk unless you encrypt it. The “Security Issues for Web-
Based Applications” section of this chapter outlines some ways you can secure your data, but
these methods are all “bolted on” and won’t have the same functionality as a native solution.
Eventually the vendors who are defining SOAP plan to add security.

Service Is in Use
Depending on how you create your application, you might find that you see a Service is
in Use error. The odd thing about this error is that it won’t have a error number, just that
single sentence. It normally means that someone crashed the application on the server and
the request timed out. Something in your code or your setup allowed the user to enter
wrong information or perform the task incorrectly. In some cases, it’s gremlins creeping into
your system, but we won’t discuss that possibility.
Of course, the more users who try to gain access to the application while the server is in this
state, the harder your job to fix it. The easiest and fastest way to fix this problem is to use
the IISRESET (or similar reset) command. This command will have the site up and running
again in about a minute.
The problem with this technique is that it kicks everyone off the server, even those who
aren’t using the errant application. This isn’t good for a system that has many users,
especially if you have database operations in progress.
You’ll notice that the server-side component in the example is an EXE, not a DLL. This
means the server-side component will show up in Windows Task Manager (see Figure 9.6).
306 Chapter 9 Moving to Web-Based Applications

You can highlight the errant component entry and click End Process. In some cases, this
will fix the problem, but you can’t count on it every time. If the error is with SOAP, the
Web server, or the processing component, you’ll likely need to restart the Web server.

Figure 9.6
The Windows Task
Manager can cure
some, but not all,
of your service-
related woes.

ActiveX Component Can’t Create Object

If you have users who like to tinker with their system, you’ll eventually see the ActiveX
component can’t create object error message shown in Figure 9.7. However, other users
will see this problem as well, because it can happen for many reasons. The most common
reasons are that the SOAP components aren’t properly registered, they became corrupted
in some way, or a problem exists with the input to the components.

Figure 9.7
This message appears
in many situations
where corruption
occurred or the com-
ponent is missing.

Sometimes all you need to do to fix this problem is click Refresh. An error in the data
stream could prevent SOAP from instantiating the component properly. In fact, it’s a good
idea to view the source coming from the server to ensure you’re getting the correct input.
Registering the SOAP components again can help in a few cases. The applications on your
machine play havoc with the registry, so it’s not unheard of for a component to lose its reg-
istry settings. The process takes moments, and you’ll know the components are registered.
 Security Issues for Web-Based Applications 307

The next thing you’ll want to try is rebooting the machine. It could be that memory is cor-
rupted and you need to clear it out. I have made more than a few fixes that didn’t show up
until after a reboot.
Finally, try reinstalling SOAP to ensure none of the components are corrupted. Windows
and other operating systems do a better job of protecting components today than they did in
times past, but corruption still occurs. Reinstalling the components won’t take too much
time, and you’ll be sure the system is clean before you begin other troubleshooting.
Sometimes it helps to use the same exact set of inputs from another machine. This check
finds server configuration and security problems in some situations. If the user’s setup fails
on the second machine, then you need to look at the server. Otherwise, there’s a problem Ch
with the first client. 9
Nothing Happens or Strange Error Message
Configuration and security errors are difficult to track under the best of circumstance. Web-
based applications will often test your skills because there are so many points where error
can occur. For example, setting the security level of your browser too high when using
unsigned components or components not marked as safe for scripting will result in no
results at all. The browser will appear to ignore your input and won’t provide a reason for
the problem. If you’re using Internet Explorer, you’ll see an extremely small message in the
lower-left corner of the status bar that simply says Error on page. This is the only notifica-
tion you receive and it disappears the second you remove your cursor from the button.
One of the reasons you need several ways to test your application is to detect problems like
this one. If you use the form-based thin client and receive an answer from the server, you
know that the local SOAP configuration and everything on the server is fine. The problem,
more often than not, is one of browser configuration. Of course, this problem could point to
something like unsigned components as well. Browsers are set to detect problems that a
desktop application might not consider.

Security Issues for Web-Based Applications

You might feel good about your company’s security measures, but it’s important to ask if that
security has a firm foundation. A recent Computer Security Institute (CSI) study conducted
with the Federal Bureau of Investigation (FBI) states that computer crime losses are on the
rise and security problems are to blame for many of those losses (http://www.gocsi.com/
prelea_000321.htm). Current losses topped $371 million; a 41% increase over the previous
CSI report. In addition, $151 million of that loss was due to theft of proprietary informa-
tion—the kind of information that a Web-based application might exchange with a server
of the Internet. Actual losses are probably higher than what I’m telling you in this section
because only 186 of the 538 respondents were actually willing to share financial information.
More losses occurred, but there wasn’t a way to know just how much more.

CSI is an outstanding resource for current security statistics and other security-related
information. You can find them at http://www.gocsi.com/.
308 Chapter 9 Moving to Web-Based Applications

Another interesting statistic is that 91% of the companies surveyed in the CSI report
detected employee abuse of Internet privileges. For example, an employee might give away
sensitive information while surfing a site for entertainment purposes or download a virus
from one of these entertainment sites. That particular statistic is important because it shows
one area where crackers hit your Web-based applications hardest. Employee abuse of the
Internet opens a door for crackers to invade your network. The cracker simply attaches a
Trojan horse virus to the employee’s machine and waits for the employee to contact the
company. In other words, employees can be unwitting aids to cracker activities.
All of these security threats come at a time when companies are making developers more
responsible for security. You’ll find that your role in ensuring applications remain safe
increases as security threats increase. Network administrators no longer know enough about
today’s applications to provide the level of security monitoring they did in the past. The best
person to fill in the gaps is the developer.
What this means to you is that you’re going to have to include more security code within
your applications. It’s no longer enough to write bug-free code that detects errors. The
code you write is now responsible for protecting the security of the data it manipulates.
Companies will require your code to detect unauthorized access and use by employees. In
short, you now share a security burden that is equal to or perhaps greater than the one born
by the network administrator.

Make sure you consider security and privacy issues at the same time during the devel-
opment process. Seventy companies are considering a standardized method for
exchanging customer data at the time of this writing. Privacy advocates are just as
certain they want to decrease the amount of customer information that companies
have to exchange. The problem is so severe that government agencies such as the
Federal Trade Commission (FTC) have gotten into the act. The FTC recently held a
workshop to look at the methods companies use to obtain, store, and exchange cus-
tomer data. If the government begins to regulate the exchange of customer data and
the means used to store that data, you can be sure that security requirements will also
increase. Building security into the portions of your application that affect privacy today
will ensure you don’t have to rush to make government-mandated changes tomorrow.
Make certain that you use flexible security and privacy code—the requirements for both
security and privacy are certain to change as the Web environment changes.

Remember from previous chapters that SOAP provides nothing in the way of security. After
reading the statistics in the CSI report, you might wonder how you’re going to create a
secure application that will protect company data with no tools for the job. The answer is
that you’ll need to secure the transport protocol and place SOAP within an envelope pro-
vided by that protocol. Secure-Hypertext Transfer Protocol (S-HTTP) provides one answer
to the question; Secure Sockets Layer (SSL) provides another. By the time you read this,
standards groups will provide yet other ways to secure your application.
 Security Issues for Web-Based Applications 309

The following sections will explore various methods you can use to secure data transmitted
by Web-based applications. We won’t look at every possible technology. I chose the tech-
nologies that I felt you would need most often and that offer the broadest base of support.
Even if you don’t use the example code in this section directly or choose a different security
technology for your company, the examples show various techniques that you can adapt to
just about any standardized security scheme.

An Overview of the Potential Security Solutions

It’s important to know the potential sources of security solutions for the problems in your
organization. The Internet as a whole has to grapple with this problem. Vendors are contin-
Ch
ually devising new technologies, but crackers seem to find holes in the new strategies just as
quickly. In short, security is a tug of war where the fastest party wins. 9

Security plays an important role in just about every business. However, security could
take on an increasingly important role as governments dictate privacy laws and begin
issuing security requirements for specific industries. For example, they recently issued
further guidelines for the financial industry. In this case, the guideline probably equates
to law because bank examiners view guidelines as suggestions that financial institutions
must follow. As you plan your security strategy, look around at other industries. Even if
the government isn’t telling you the minimum acceptable security limits for your busi-
ness today, you might find those limitations in place tomorrow.

The following sections will look at some potential security solutions for your SOAP applica-
tion. All of these methods are standards that Internet vendors currently use for Web sites
and other users. Of course, the best solutions will arrive when SOAP finally has security of
its own.

S/MIME
Securing e-mail transfers, especially if you use them with a SOAP application, is important.
E-mail is not only exposed during the time of transfer, but also while it sits on the server
and then in the user’s Inbox as well. A consortium of vendors, including Microsoft, Banyan,
VeriSign, ConnectSoft, QUALCOMM, Frontier Technologies, Network Computing
Devices, FTP Software, Wollongong, SecureWare, and Lotus are promoting this standard.
RSA Data Security, Inc. originally developed Standard MIME (S/MIME) as a method for
developers to create message transfer agents (MTAs) that used compatible encryption tech-
nology. Essentially, this means that if someone sends you a message using a Lotus product,
you can read it with your Banyan product. S/MIME is based on the popular Internet MIME
standard (RFC1521).

You can find out about standard MIME at http://www.oac.uci.edu/indiv/

ehood/MIME/. In addition, you can find an entire list of S/MIME-specific resources at
http://www.rsasecurity.com/standards/smime/.
310 Chapter 9 Moving to Web-Based Applications

S/MIME represents one of the better ways to transfer data if you plan on using a one-way
communication scenario. For example, using an S/MIME transfer makes sense if you plan to
create a SOAP application for surveys. Unfortunately, until SOAP toolkit vendors begin to
support MIME and S/MIME as part of an HTML transfer, you won’t be able to use this
technology for two-way communication. It’s possible to do so, and vendors are looking at
ways to implement SOAP with attachments—essentially a form of MIME as I write this.

Solutions That Require Two-Way Communication

One of the problems that I keep mentioning about working with the Internet is the lack of a
secure connection. When it comes to security, however, the lack of a connection could turn
into a significant problem. Many security protocols in use today require a two-way connec-
tion that remains intact long enough to exchange validation information and then the data,
which is a minimum of four trips from client to server and back in most cases.
Security standards such as SSH, SSL, TLS, IPSec, Kerberos, Kracken, SRP, and PAK all
share the same need for two-way communication. Using these standards to secure your
online data ties you to transport protocols, such as HTTP, which support two-way commu-
nication. In addition, you’ll find that the client and server must exist at the same time—one
of the problems people are trying to solve when using distributed applications. Developers
want to work with disconnected applications that can rely on protocols such as SNMP in
addition to HTTP.
Although these two-way communication techniques work fine for a LAN, you’ll find them
cumbersome for a Web-based application. This is one of the reasons that SOAP should have
a built-in security mechanism that is more in line with its purpose.

Payload Protection
Sometimes the best security solution is the one that requires the least effort to implement.
Payload protection falls into the easy-to-implement category because you don’t need a lot
of special software to do it. Using the MIME message technique we discussed in the
“Understanding SOAP Attachments” section of Chapter 2 allows you to send a SOAP
message and an attachment. The attachment could contain the sensitive data.
Of course, you need to provide some type of protection for the payload. It’s important to
protect the sensitive data in such a way that only those with the proper credentials can
access it. The problem with most security solutions is that the user has to go through a lot
of bother to use them. That’s why you still don’t see much encryption for e-mail messages.
The user has to go out, get a certificate, and then go through the trouble of encrypting
the message.
Alternative encryption products could make payload protection both safe and easy. For exam-
ple, Masker (http://www.masker.de/) allows you to encrypt data in such a way that it’s still
viewable using a double-click. The only difference from a user perspective is that the pro-
gram now asks for a password before Master will grant entry. Your SOAP application could
use a MIME message to transfer the Masker-encrypted payload from one location to another.
Masker relies on the same RC4 encryption algorithm used by encryption technologies such
 Security Issues for Web-Based Applications 311

as SSL, which is reasonably safe, but not completely secure. In other words, you could trust
it to transfer sensitive, but not critical data. We’ll look at the techniques for using Masker in
Appendix C, “Third-Party Tool Reference.”
The industry term for this type of encryption is a carrier file. An application creates the
carrier file around the data file. The carrier file normally provides the required encryption.
(Some types of carrier file products merely provide data hiding, not data encryption.) The
application runs in the background on the client machine and provides automatic decryption
services as needed. In short, carrier files provide invisible data protection.
Carrier files aren’t new. Products such as the wbStego Steganography Tool (http://
Ch
wbstego.cjb.net/) have successfully hidden data files within bitmaps, text files, HTML files,
and even PDF files for quite some time now. Encryption is actually an optional part of this 9
program because it focuses on data hiding, rather than data encryption.

Using SSL
One of the currently viable solutions for the security problem is SSL. Many developers find
that they need to test SSL with standard Web pages before they use it with SOAP. Getting
a server set up correctly seems to be a problem (as evidenced by the many SSL-related
messages on the various newsgroups).
To use SSL, you’ll need certificates for both client and server either issued by a known cen-
tral authority such as VeriSign or generated using a certificate server such as the one found
with Windows 2000. After you have the certificates installed on both machines, make sure
you attach the certificates to the root authority of the machine. Otherwise, the certificates
will exist, but you won’t be able to use them properly.
Always make sure to test your code using a clear connection before you add security.
Otherwise, simple communication problems might appear as security problems. After you
debug the application, you can modify the SOAP initialization call as shown here:
Client.mssoapinit _
“https://winserver/soapexamples/ComputerName/CompNameProc.WSDL”, _
“CompNameProc”, _
“NameValuesProcSoapPort”

Notice that the only change is to convert “http” to “https” in the second line. The use of
SSL should be transparent to the end user and nearly so to you as the developer. If you find
that SSL presents a lot of problems, make sure you have the Web server configured properly
and that you install the certificates correctly. These two issues seem to create the vast major-
ity of problems for developers. As previously mentioned, keep running tests with standard
Web pages until you are certain that you have SSL running on the Web server.

Using IBM Web Services Toolkit

The Web Services Toolkit (http://www.alphaworks.ibm.com/tech/webservicestoolkit)
is one of the few that currently support both encryption and digital signatures. Like the
Microsoft SOAP Toolkit, the Web Services Toolkit supports many of the latest SOAP inno-
vations, including WSDL files. The Web Services Toolkit goes further, however, and
312 Chapter 9 Moving to Web-Based Applications

supports some vertical market standards such as Trading Partner Agreements Markup
Language (tpaML). You’ll find versions of the Web Services Toolkit for Windows NT (with
SP6 installed), Windows 2000, and Linux.
Given that the Web Services Toolkit works on both Linux and Windows, you might find it
the optimal solution for applications that have high security needs. Unlike other solutions,
this one has security built in, not bolted on as an afterthought. Integrated security tends to
provide better protection than protocols such as SSL.
Adding encryption to the Web Services Toolkit does modify the SOAP header. The messages
sent by the Web Services Toolkit have to indicate that encryption is in place so that the
receiver knows to decrypt the message. The extension that the IBM solution uses might not
be compatible with other solutions when they appear on the market. It is my hope that the
vendors will get together and create a standardized method for adding encryption.
A couple of special requirements exist when using the Web Services Toolkit. Make sure you
look at the requirements list for this product. At the time of this writing, you needed a full
installation of Java to work with the product. In fact, the Web Services Toolkit includes
Universal Description, Discovery, and Integration for Java (UDDI4J) to make it easier to
advertise services. The examples that I used also required the addition of a plug-in for the
browser, which you might not always need.
Note that this product runs on WebSphere Application or Apache Tomcat servers. If you’re
already running one of these two servers, upgrading support for the Web Services Toolkit
won’t present problems. On the other hand, if you’re running IIS or another Web server
product, using the Web Services Toolkit could become problematic or even impossible.
Fortunately, the Web Services Toolkit does include an embedded version of the WebSphere
Application Server so that you can try it before you make a commitment to this product.
Problems aside, if you absolutely have to have the very best security right now, the Web
Services Toolkit is probably your best option. Make sure you purchase a compatible Web
server before you create production applications because the embedded server is only useful
for test scenarios.

Quick Fixes for Memory and Other Resource

Problems
Web-based applications are one of the newest application types for any developer. The
Internet has radically changed how developers create applications and how users interact
with those applications. Developers no longer tie the Web-based application to the desktop
machine, although you can certainly use it in that role. A developer needs to consider more
than one client type now—everything from a desktop machine, to a laptop, to a notebook,
and even personal digital assistants (PDAs). It’s little wonder, then, that there’s actually more
room for error in the new world of the Internet, not less.
 Quick Fixes for Memory and Other Resource Problems 313

Unlike the desktop or LAN application, the world of the Internet is largely unexplored now.
Yes, you can find a lot of code, but most of it won’t run with anything else because standards
are few and reliable code is in short supply. Every developer who creates an application for this
environment is a pioneer, and like all pioneers, he is subject to the vagaries of exploration. You
might find that the perfect application you created works just fine on the company laptop
machines, begins to falter on a notebook, and finally fails when it comes to PDA support. The
use of multiple clients means that testing for a Web-based application is more extreme than
anything created for the desktop.
Internet applications also have a worldwide appeal. You can no longer count on a particular
set of biases when it comes to the user interface. Not only do you have language to consider, Ch
but also every nationality has different ways of working with computer applications. In some
cases, you might have to write the application for the lowest-common denominator and
9
allow user customization to account for differences in tastes.
All of these issues aside, the Web-based application is inherently different from anything
you might have created for the desktop in other ways. For example, you need to consider
more interfaces and data conversions. The data that you store in SQL Server no longer
undergoes a single transition from the DBMS to the client application. Data undergoes
several levels of change as components and services modify it to fit within the HTML
specification for browsers and browser-like applications. Of course, this also means
tailoring data types for SOAP.
The following sections can’t help you locate and kill every potential problem that you’ll run
into with Web-based applications. Consider these sections as a starting point, a list of the
most common problems that you’ll run into as you work with the new Internet applications
your company will demand. We’ll consider the most likely issues that every developer will
run into.

Component Interactions
Component interactions can take on a new meaning with Web-based applications. For
example, the applications in this chapter rely on multiple components that all have to com-
municate correctly to send the data to the client application. These components appear on
more than one machine and might not appear in the same place for all implementations.
The processing component appears on the client machine when used with a thick client and
the server when used with a thin client. One form of the thin client even relies on script
code, with all of the problems that scripts can entail.
The loss of connection can create situations where the client and server can’t communicate
well. Because of the way HTML handles keep alives, the client could lose the context it estab-
lishes with the server before the client completes whatever task needs to be accomplished. The
point is that you must treat every query as a new call, even if the Web server keeps the original
connection alive. (This is unlikely with SOAP, but could happen in rare instances.)
314 Chapter 9 Moving to Web-Based Applications

Scripting also presents other problems. Creating an instance of an object using a script is
seldom the same as creating an instance of the same object with a desktop application because
of the scripting environment. Consequently, you might find that components won’t work at
all with your script or might behave differently than you expect them to. You can get around
this problem by testing a component thoroughly during development. Make sure you include
scripting tests as part of your test suite. It’s also important to make sure the component
behaves the same no matter what type of desktop application or script is calling it.
Developers design many components with callbacks in mind. The client makes an initial
query, and then the component uses a callback to obtain additional information from the
client. The callback actually issues a request to a special section of the client code to obtain
the additional information or data processing. Because Web-based applications are unreli-
able when it comes to communication, it’s important to design the component to rely on
the client’s initial query without callbacks.

Benefits of the Script Debugger

Scripts are a requirement when working with Web-based applications. In fact, you’ll proba-
bly need a combination of client-side and server-side scripts to make the application fully
functional. Client-side scripts normally don’t present much of a problem when it comes to
error handling. Either the script works or it doesn’t. Although the error code the user gets
might seem somewhat cryptic, it’s usually easy to locate the problem and fix it because
client-side scripts are small and usually limited in scope.
Server-side scripts present other kinds of problems. Most scripting languages don’t provide
anything in the way of error handling. In addition, passing the error information along to
the client might be difficult without a lot of additional coding. Finally, unlike standard desk-
top applications or components, scripting languages provide little access to the event log,
denying the developer of even this potential method of recording errors. In summary, the
developer has to be exceptionally careful when creating the server-side script. Providing
range and other non-error producing checks will help reduce problems. If a range or other
easily detectable problem is encountered, the script can always provide client feedback in
the form of a special Web page.
Fortunately, Windows 2000 provides rudimentary script debugging. You need to install this
feature separately using the Windows Component Wizard accessed using the Add/Remove
Programs applet in the Control Panel. Figure 9.8 shows the required entry. After you add
this feature to your development platform, you’ll always be able to debug scripts. You need
to install this feature on both the client and the server; the Script Debugger doesn’t provide
remote debugging capability.
If you have full Visual InterDev support installed on the client and the correct debugging fea-
tures installed on the server, you can perform remote debugging. When a script error occurs
on either the client or server, you’ll see a dialog box asking if you want to start debugging the
application. A yes answer will start Visual InterDev and allow you to perform full debugging
of the application. This support is supposed to be even better with Visual Studio .NET.
 Quick Fixes for Memory and Other Resource Problems 315

Figure 9.8
Windows provides a
script debugger you
can use to locate
problems with your
script code.

Ch

Developers who rely on IIS will want to configure it to provide debugging support. Right-
click the Web site in question and select Properties. You’ll see a Web Site Properties dialog
box. Select the Home Directory tab and click Configuration. You’ll see an Application
Configuration dialog box. Select the App Debugging tab and you’ll see a dialog box like the
one shown in Figure 9.9. This is where you’ll choose the type of remote debugging that IIS
will support. Other Web servers are likely to support debugging in other ways—check your
vendor documentation for details.

Figure 9.9
Verify that IIS is con-
figured to provide full
debugging support.
316 Chapter 9 Moving to Web-Based Applications

IIS also allows you to set debugging at the directory level. This is actually a safer option
for Web servers exposed for the Internet. You can set most directories to disallow debug-
ging and add it to those few directories that you’re using for application development.
The location of the Configuration button differs by object type. For example, you’ll find it
on the Virtual Directory tab of a virtual directory.

Human Language Support

One of the biggest problems you’ll run into when working with a Web-based application is
the issue of human language support. The Internet isn’t a closed environment. If you plan to
make a Web-based application open to the public through the Internet, you’ll need to pro-
vide support for more than one language.
In many cases, the problem isn’t one of translating the content of your Web site. There are
many well-understood methods for translating text from one language to another. The main
problem is one of application usage. For example, the captions on your application will need
to change to support the other languages. You could place the required information in a
database that the application could download as needed. The most recently used languages
could be stored in a local database or even the registry.
You need to consider broader issues, however. For example, the layout of your application
must be flexible enough to accommodate the way that people normally work or users will
complain that the application is difficult to use. This means allowing users to configure the
display. Although this isn’t a problem with most desktop applications, it’s a little more diffi-
cult with a Web-based application because you don’t have control over the position of the
various components. In most cases, the best that you can do is store the user preferences in a
cookie, then use the contents of the cookie to configure the display as data downloads from
the server.

ASP and Component Communication

One of the ultimate problems with scripting is that you lose the direct connection between
the client and the server. The SOAP application might generate an error, but unless you
design the script that formats the Web page to pass that error along, the user might never
see it. Because a Web-based application normally has a direct connection (you wouldn’t
create a disconnected application in many cases), it’s important to pass any component
errors to the client.
Unfortunately, simply passing the error along isn’t enough in most cases. The problem is
that the user won’t know where the problem occurred unless you provide additional infor-
mation. In most cases, the user will assume that an error has occurred locally, unless you
make it clear that the problem occurred within the component. Fortunately, SOAP helps in
this regard by making at least its error messages server oriented. Unfortunately, the SOAP
messages are still ambiguous at times.
 Case Study 317

Of course, complex applications might have several layers of component calls. Therefore, it
might be tempting to add code so the user sees the precise location of the problem. Rather
than build overly complex components that do little for the user, it’s better to simply report
the error as a component error, then tell the user to contact the network administrator
about the problem. An event log entry will allow the network administrator to determine
the precise cause of the problem and take steps to fix it.

Component Doesn’t Support Locale Error

Increasingly, desktop applications are required to support multiple languages, but it’s still possi-
ble to write a custom application and not worry about the locale. You can’t say the same of a
Ch
Web-based application. By its very nature, a Web-based application requires multiple language
support. Of course, this presents a problem for the developer because he might not know which 9
locales an off-the-shelf component will support, if it supports more than one language at all.
Visual Basic applications, including components, will generate a 477 error code when a
component called by the application doesn’t support the current local setting. When you
detect this error in your Web-based application, you’ll normally have three courses of
action. First, you can ask the user if some other language will work, in which case you’ll
have to try the call again with the new locale. Second, you can try to get around the prob-
lem by attempting to access the component with another dialect of the same language. For
example, several locales support French. The resulting text might not be perfect, but it
should be readable. Third, you can register a failure with the user and present any data you
were able to retrieve before the error occurred.
The 477 error code doesn’t always occur when a component lacks support for a particular lan-
guage. The LCID table has had several versions over the years, therefore the one used by your
component might be out of date. Getting a newer version of the component will usually help.
If that isn’t possible, you might need to find an alternative LCID that the component does
support. Again, this isn’t a perfect solution, but it might be the only choice in some cases.

Case Study
Web-based application development may seem like a pipe dream that will never come to
fruition for some people. However, some companies have already seen results from using
SOAP to re-create their desktop applications as Web-based applications. In fact, the move
from the desktop has re-created some companies to the point the corporate structure is no
longer the same. Consider the work done by Etelos.com. I had the pleasure of talking with
Ahmad Baitalmal about their new application, Referral Survey.

An Overview of the Problem

Etelos.com realized early during a market downturn that they were in the same predicament
as many other companies. They had a great Web site with some loyal customers, but the
Web site wouldn’t generate enough income to keep the company afloat. Etelos.com provides
318 Chapter 9 Moving to Web-Based Applications

a referral service named Referral Survey that sales people of other companies use to become
more productive and eventually more profitable as well.
Referral Survey is a special type of referral service. Some organizations have a large cus-
tomer base that is hard to track. In many cases, they don’t take full advantage of their
customer base as a resource for new business. Prompting each customer to provide com-
ments and a referral generates a lot leads. These are qualified leads because customers
who know what services the company provides generate them. Referral Survey allows
companies to automate the process of generating the customer requests and collecting
the comments and sales leads.
The upper management of the client companies saw the value of Referral Survey. Salespeople
with better tools who could work more efficiently are generally more profitable for the com-
pany as well. However, the sales people were resistant to using the new tool. The Etelos.com
staff tried to train them and make them more aware of the potential of Referral Survey, but
the results were disappointing. The client companies didn’t realize the benefits they had
originally sought.
The problem is one of, “we don’t do things that way here.” The sales staff didn’t want to
interact with previous customers directly—they wanted to maintain the old way of working
with them. It’s a recurring problem in many companies. Education and making people aware
of new tools that can help them only goes so far. Of course, it’s easy to blame the people in
this situation, but there are other factors to consider. For example, the new tools required a
conscious effort to use and weren’t as transparent as they could be. The sales force had to
work too hard to make the new tools work.

The Solution
The Etelos.com programming staff looked at the whole referral system provided by Referral
Survey. They looked for ways to automate every important feature of their product. Once
they found some areas they could automate, Etelos.com went back to the clients with their
solution. The new system overcame many of the negatives that the client sales people com-
plained about, yet kept all of the positive features the tool could provide.
In its current state, Etelos.com looks at the sales process at an organization and identifies
where an “automated” survey will enhance customer relations. The client’s system calls a
method on Etelos.com’s system to send a survey, collect comments, and prompt for a referral.
There are no sales people to train on using this tool, which reduces costs because only a sin-
gle administrator has to know about the Etelos.com system and configure it. In short, the
new system completely automates a task so that all the client’s sales staff needs to do is make
calls to potential customers.
 Case Study 319

Now that everyone was on board, the Etelos.com programming staff faced an even more
difficult task. All of the automation they added to the application meant creating better
integration with the client systems. Unfortunately, every client had a different system and
the Etelos.com programming staff faced the prospect of creating unique interfaces for
each client.

The SOAP Connection

The Etelos.com staff decided to use SOAP to create a Web service interface for their appli-
cation. Since their system already relied on Microsoft technology, making the transition
using the Microsoft SOAP toolkit was relatively painless. Within a month, Etelos.com Ch
created a solution to expose their existing services. They had planned to spend three months
on the project using DCOM, so the time they saved was a nice plus of using SOAP. Not 9
only did they save time, but also the programming staff feels they have far fewer headaches
implementing the SOAP solution.
The changes the Etelos.com staff made delighted clients that already knew about XML and
SOAP. The changes made it easier for the client to interact with the server. In addition,
changes to the client’s hardware wouldn’t affect their ability to work with Etelos.com
because SOAP effectively isolated them from the changes.
Even the configuration interface uses SOAP. Using ActiveX to build the Administration UI
and SOAP to communicate with the Etelos.com service allows the client to configure their
system directly. This makes life easier for the client’s administrator who no longer needs to
call Etelos.com to make changes, and reduces support costs for the Etelos.com who no
longer has to staff as many people to make configuration changes.

A Negative to Consider
One of the problems that the Etelos.com staff has with SOAP is that it doesn’t generate
events. This means that a client must poll for new referrals, rather than get them as soon as
they become available. Polling wastes valuable network bandwidth, so using events would
have made the system more efficient as well. Etelos.com considers this a minor problem
because of the other benefits they received.
 CHAPTER
10
Working with PDAs
In this chapter
Special Needs for PDAs 323
Getting SOAP for Your PDA 326
Updating the Complex Type Example 329
Updating the Computer Name Example 329
Addressing PDA Display Issues 338
Beyond PDAs to Telephones 341
Understanding PDA Security Issues 341
Troubleshooting 343
322 Chapter 10 Working with PDAs

The world is becoming more mobile all the time. It’s nothing to see someone talking over a
cell phone today, and just a few years ago, it was a novelty. When I first saw a Personal
Digital Assistant (PDA) in 1998, I thought they might be a passing fad or a device of limited
use. Today that vision has changed significantly. Developers create applications for PDAs
now that many people would have thought impossible even a year ago.
SOAP is an excellent protocol for PDAs because the people who use them spend most of
their time on the road. PDAs always perform remote communications with the company
network because vendors don’t design them to communicate in any other way. In short, if
you own a PDA, you need good remote communications of the sort that SOAP can provide.
Obviously, PDAs have special programming requirements. They don’t have a hard drive in
the normal sense of the word, lack a high-speed processor, and even memory is at a premium.
The first section of this chapter, “Special Needs for PDAs,” examines the special program-
ming required for PDAs. You’ll find that you need to jump over some high hurdles to make
some types of applications work on a PDA, even with the help of SOAP.
At the time of this writing, Microsoft doesn’t provide a toolkit that supports SOAP on a
PDA, not even for Windows CE machines. As a result, we’ll examine several other SOAP
toolkit choices in the “Getting SOAP for Your PDA” section of this chapter. We’ll look at
choices for several popular PDA operating systems.
The third (“Updating the Complex Type Example”) and fourth (“Updating the Computer
Name Example”) sections of this chapter will show how to create SOAP applications for a
PDA. Both of these examples rely on components that we’ve already worked with in the
book. We’ll create examples for the PDA that require no server-side coding changes at all.
That’s one of the benefits of working with SOAP—cross-platform compatibility. Consider
these examples as the proof-of-concept sections of the book. Certainly, developing examples
that will work on both a PDA and a desktop with equal functionality is a challenge that
other technologies would be hard-pressed to match.
The fifth section of the chapter, “Addressing PDA Display Issues,” will discuss some of the
problems that you’ll encounter with PDA displays. The most obvious problem is the lack
of screen real estate to display anything. Developers who work in the world of PDAs will
quickly find they need to provide ingenious solutions to the display problem.
You won’t work with PDAs forever. Cell phones are certain to follow as a device that devel-
opers will learn to hate as they create new distributed applications for even smaller devices.
I won’t provide you with an actual example in the “Beyond PDAs to Telephones” section
of the chapter, mainly because SOAP toolkits aren’t available for this platform yet. We’ll
discuss many of the problems you’ll face as you move to this new platform.
The seventh section of the chapter, “Understanding PDA Security Issues,” discusses security
problems that are specific to PDA development. These security problems are in addition to
those you’ll normally find with any SOAP application. It’s important to understand that ven-
dors are working on these problems even as I write this chapter, but the solutions will likely
take time to fix.
 Special Needs for PDAs 323

Finally, we’ll look at some troubleshooting techniques for PDA applications based on the
content of the entire chapter. You’ll learn some of the ins and outs of working with a PDA
in a development environment. This section will tie everything together in a neat package
for you as you begin writing PDA applications of your own.

PDA development is moving so fast that it’s hard for anyone to keep up. Figuring out
where PDAs will go in the future is even harder. Fortunately, someone is already per-
forming the research required to show you current trends in PDA development. Learn
more about the future of PDAs at http://researchportal.com/. The Research
Port site does contain information about many different aspects of computers today,
but you’ll find that they cover mobile computing topics especially well.

Special Needs for PDAs

There’s no free lunch—I often wish I had been the first one to say that because it’s so true. Ch
Companies that want to gain the advantages of using PDAs also have to decide how to han-
dle the special needs of these devices. A PDA isn’t a single-purpose device like a radio, but it 10
isn’t a full-fledge computer either; it’s somewhere in-between.
After spending some time working with several PDAs for this chapter, it’s apparent that
SOAP is even less ready for prime time when it comes to these devices than it is for the
desktop. (The test machines included a Casio Cassiopeia and a Palm VII.) This makes sense
because software of this sort normally appears on the desktop first. However, it also means
that you need to consider your PDA development plans carefully because of the many pit-
falls. In all cases, you’ll want to build a desktop version of your application before you
attempt to create one for your favorite PDA.
Fortunately, you can create a SOAP application for your favorite PDA; it just takes a little more
planning. The following sections examine the special needs of PDAs. It’s important to note that
most of these special needs are in addition to what you must consider for a desktop application.

The Case for PDAs

Developing SOAP applications for PDAs will require considerable work. Many developers are
unused to working with devices that have small screens; the memory limitations are problematic
at best, and there’s a limit to the number of available development tools. However, the need for
PDA development is strong.
Consider the case of Sears. They recently purchased 15,000 Palm PDAs for their business.
The deal is worth between $20 million and $25 million. That’s a lot of PDA power for their
staff. Each PDA is equipped with a built-in bar-code scanner and a wireless modem. The
company plans to use these new devices for inventory management, price changes, and mer-
chandise pickups. Some developer has a large programming task in the works as I write this.
You can bet that such a serious investment comes with an equally serious need to affect the
bottom line. In short, PDAs are becoming mainline systems for many different tasks.
324 Chapter 10 Working with PDAs

Companies often cite two main reasons for switching to PDAs after using other devices. The
first reason is that PDAs cost less to buy, operate, and maintain than many other devices. A PDA
equipped with the right add-on devices can perform a myriad of tasks in an intelligent manner.
The second main reason is ease of use. PDAs have a limited number of buttons on them and
the functions of each button are easy to understand. The user writes on the screen as he
would using pen and paper—the PDA uses handwriting recognition to convert the hand-
written information into text.
Sears might be looking toward the future as well. A customer could come into the store
with a PDA, beam the information to a sale clerk’s PDA, and get his merchandise faster
than ever before. Of course, this use is in the future; most shoppers today don’t place their
orders in a PDA. The point is that Sears and other companies like K-Mart are already
planning for this eventuality.
Unlike older, single-function devices, PDAs are completely programmable. This means an
investment in hardware today won’t become an albatross tomorrow. Companies can extend
the life of an investment by using the same PDA in more than one way. As the PDAs age,
they’ll handle applications with lower programming requirements.

Special Add-ons
Most vendors design PDAs as electronic versions of the calendar, address book, and personal
note taker. Early versions of these products didn’t include the mini-word processors and
spreadsheets you’ll find in modern versions. In fact, today you can extend many PDAs to
double as cameras, scanners, and other devices with special add-ons.
The PDA isn’t exactly a standard device. There are many hardware implementations, more
than a few operating systems, and even different capabilities to consider. When users start
adding features to their PDA, you might find that it’s nearly impossible to determine which
features you can rely on finding. In short, standardization within the company is essential,
even if there’s chaos outside.
These special add-ons can also work to your advantage. Imagine creating an application to
work with one of the camera attachments for a PDA. Each picture is automatically trans-
ferred to a remote processing center as the photographer takes pictures. SOAP could make
this task relatively easy and automatic. The pictures would be ready for viewing by the time
the photographer reaches the home office.
In summary, special PDA add-ons present problems because they create a non-standard
programming environment. On the other hand, these add-ons can create new productivity
situations where a developer can provide functionality that no one has ever seen before.
The optimum setup is to standardize features required to make SOAP work, such as the
type of network interface card (NIC). On the other hand, it’s important to consider spe-
cialization. Adding a camera to a PDA turns it into a direct image transfer device. PDAs
provide the means to extend what computers can do as long as you configure them with
application connectivity in mind.
 Special Needs for PDAs 325

Networking
SOAP relies on a connection between the client and the server. It’s easy to create a connec-
tion when you’re working with a desktop machine. If you can’t create a direct connection
using a LAN, there are always alternatives such as using dial-up support or a wireless net-
work connection. SOAP makes communications between a server and desktop machine easy
because of the many ways to create the connection.
PDAs might not offer much in the way of network connections. In many cases, the PDA
vendor will offer a method to synchronize the PDA with the desktop. The synchronization
process works with static data and won’t provide you with a live connection to the network.
It helps to have a wireless network setup when working with a PDA because many vendors
design PDAs to use this connection type. I was able to find a third-party product for con-
necting a Pocket PC directly to the network using a special NIC.
Every PDA that I looked at does provide some type of modem support, but the modem is usu-
ally an add-on and doesn’t come with the device. Again, you’ll need to standardize the kind of
modem support you need for your application. You’ll also want to be sure that you can create Ch
the desired connection and view material using a browser if you decide to go that route. 10
The bottom line is that the original purpose of a PDA conflicts with the ways that some
people use them today. Vendors designed the PDA to provide electronic versions of calen-
dars and address books. Yes, you can run SOAP on a PDA, but only if you have the required
live network connection. Obtaining that connection can prove difficult to say the least.
Vendors seem to know that this is an issue and at least some of them are working on ways
to overcome these problems.

Operating System
You probably know about the major PDA operating systems on the market today. Palm is one
of the most common PDAs, and it has its own special operating system. Windows CE (also
known as the Pocket PC) is another favorite because it looks and acts much like Windows for
the desktop. What you might not realize is that the operating system the PDA package says it
uses isn’t the operating system you’ll get. Small variations between machines make the differ-
ence between a SOAP toolkit that works and one that won’t even install.
The operating system also determines what type of client you can run. Palm systems offer
ease of use and low cost. However, they also offer fewer features and an operating system
that is less capable than Windows CE. If you plan to target the Palm, then you should prob-
ably target Java as your programming language and plan to reduce the number of advanced
application features.
Windows CE allows multiple levels of development. The operating system provides a subset
of the features found in the Windows API. This means that you can use the programming
techniques you learned in the past if you’re already a Windows developer. There’s even
a special toolkit for working with Windows CE devices (http://www.microsoft.com/
catalog/display.asp?subid=22&site=763&pg=1). This toolkit includes a Windows CE emu-
lator that allows you to test your application without a Windows CE device, along with
add-ons for Visual Basic that make development easier.
326 Chapter 10 Working with PDAs

Because Windows CE also contains Internet Explorer, you can interact with it using a
browser application. Be warned, though, that the version of Internet Explorer that ships
with Windows CE doesn’t include full scripting support. You can use JScript (Microsoft’s
form of JavaScript), but not VBScript (see “Internet Programming with Windows CE”
http://www.microsoft.com/mind/0599/webce/webce.htm for details). Windows CE will
simply ignore script commands it doesn’t understand in the Hypertext Markup Language
(HTML) file. In addition to JScript, you also have access to some Internet Explorer func-
tions such as Alert(), and you can use standard HTML tags. The Windows CE version of
Internet Explorer will also work with Java applets. This means that you can create complex
browser applications that don’t rely on VBScript.
As PDAs increase in complexity, the operating systems that support them will as well.
This means that SOAP toolkits of the future should provide robust capabilities for PDAs
and even cellular telephones. Consider the capabilities of today a mere shadow of what
lies ahead.

Getting SOAP for Your PDA

If you’re the lucky owner of a Windows CE machine (also known as the Pocket PC), getting
a SOAP toolkit is relatively painless. Palm owners are next on the list. At least one vendor is
developing a solution for this platform. None of the other PDA choices on the market has a
SOAP toolkit available as of this writing. However, given the newness of this technology,
you can expect other vendors to provide SOAP toolkit offerings for other platforms eventu-
ally. The SOAP::Lite site (http://www.soaplite.com/) contains a section of SOAP toolkit
links you can check periodically for new additions. This list tells which toolkits will work
with PDAs.

According to a recent Research Portal study, developers are most likely to favor hand-
held devices that use the Microsoft (32%) or Palm (27%) operating system. However, if
you consider that these top two operating system choices only control about 59% of
the market, it’s plain that developing for any given platform could be risky. There are
two ways to get around this problem. First, ensure your company standardizes on a
single handheld device operating system if possible. Second, make sure you write most
of the processing code for a PDA application to run on the server, rather than the
client. This allows you to write for multiple PDA operating systems with less effort.

Despite the long list of SOAP toolkits you find on the SOAP::Lite site, most aren’t ready for
prime time. The vast majority are in beta or not in a released state at all. The choices of usable
toolkits for SOAP are extremely limited now, but you should see more choices as SOAP
becomes entrenched within the corporate environment. The bottom line is that you not only
need to find a SOAP toolkit for your PDA, but you need to find one that’s fully functional.
 Getting SOAP for Your PDA 327

It’s impossible to know at this point just how many SOAP-related specifications will even-
tually appear on the horizon. One of the best places to learn about new specifications is
the XML Web Service Specifications page at GotDotNet (http://www.gotdotnet.com/
team/xml_wsspecs/default.aspx). Three new specifications recently appeared on
the standards groups agenda, including SOAP Routing Protocol (SOAP-RP), Direct Internet
Message Encapsulation (DIME), and XLANG. Each of these three specifications is so new
that not much is written about them yet. However, Microsoft uses XLANG (http://www.
gotdotnet.com/team/xml_wsspecs/xlang-c/default.htm) with their BizTalk
Server product (see Appendix B, “Microsoft Biztalk and SOAP”). XLANG will allow
developers to model business processes using a standardized syntax. SOAP-RP (http://
www.gotdotnet.com/team/xml_wsspecs/soap-rp/ default.html) makes it
easier to move data using SOAP over transports such as TCP, UDP, and HTTP in one-way,
request/response, and peer-to-peer scenarios. DIME (http://www.gotdotnet.com/
team/xml_wsspecs/ dime/default.htm) is used to package data using a binary
format in a form called payloads. Vendors are now trying to make all of these variations
on a theme fit within a framework (http://www.w3.org/2001/03/WSWS-popa/
paper51). The idea of a framework is to show how the pieces fit together into a cohesive
whole. You can monitor progress on the framework as well as other XML projects at Ch
http://www.w3.org/2001/04/wsws-proceedings/ibm-ms-framework/. You’ll
also want to check the discussion groups for SOAP-RP (http://discuss.develop.
10
com/ soap-rp.html), DIME (http://discuss.develop.com/dime.html),
and XLANG (http://discuss.develop.com/xlang.html).

The following paragraphs provide a quick overview of some of the better SOAP toolkit choices
for PDAs today. This list isn’t exhaustive or even partially complete. I chose to concentrate on
those toolkits that are close enough to completion that you can use them for development today.

pocketSOAP
This is the best choice if you own a Pocket PC. The same developer, Simon Fell, produces
Simon’s Soap Server Services for COM (4S4C) (see Appendix C, “Third-Party Tool
Reference,” for details) and pocketSOAP (http://www.pocketsoap.com/). We used the desk-
top version of pocketSOAP in Chapter 8, “Providing Remote Database Access.” The Pocket
PC version is just as easy to install and works just as well within the confines of the Pocket
PC’s feature set. In some cases, you can move code directly from your desktop to the
Windows CE machine with few changes. This is especially true if you use scripts within a
Web page as we will for the examples in this chapter.
Although pocketSOAP is still in beta, you’ll find that most applications work with few prob-
lems. The only major problem that I experienced during testing was an occasional HTTP
timeout error. The developer has promised to keep working on the kinks, so you’ll likely
find this product an optimum choice for PDA development.
The only caveat when using pocketSOAP is that you need to create the message by hand. It
doesn’t support a high-level API like the Microsoft SOAP Toolkit. However, this actually
turned out to be beneficial when working with more than one platform, as we’ll see in the
examples. The bottom line is that you need to be prepared to spend a little more time coding
when working with pocketSOAP, but the outcome is well worth the effort.
328 Chapter 10 Working with PDAs

IdooXoap
IdooXoap (http://www.idoox.com/idooxoap.html) is a fully developed desktop product. It
currently comes in Java and C++ flavors that many developers will find useful for both Web-
based and desktop applications. The vendor is currently working on a version of IdooXoap
for the Pocket PC. Unfortunately, the PDA product wasn’t ready at the time of writing.
One of the advantages of IdooXoap is that it provides cross-platform support. The server
part of the product runs on Linux or Unix. Eventually it will run on Windows without mod-
ification as well. (There’s a downloadable ISAPI package for IdooXoap on the Web site, but
you need to put the package together before you can use it.) The client runs on Linux,
Unix, and all flavors of Windows (except Windows CE).

IdooXoap provides the best support for multi-part MIME messages. The support is only
available on Linux and Unix machines now, but might become available on other plat-
forms. Compatibility issues flaw the MIME support, but the vendor might have them
fixed by the time you read this. You shouldn’t count on MIME support for PDAs any-
time soon because of resource limitations. However, given the progress this particular
vendor has made, the possibility of MIME support on a PDA does exist.

A major goal of IdooXoap is to provide the same programming experience no matter which
platform you’re using. Like pocketSOAP, this means that you should be able to move at least
some of your code between platforms. However, the Java version is more likely to provide
seamless support.
IdooXoap also provides special support for the Tomcat server. You can download versions
with or without Enterprise Java Beans (EJB). It was interesting to not that several of the
specialty downloads include notes about daily product builds—an indicator of the volatility
of the SOAP toolkit market.

kSOAP
Most of the SOAP development efforts for the PDA currently center on the Pocket PC. Many
developers say that the Pocket PC provides more resources for them to use in creating SOAP
applications and that the Palm is somewhat limited. Given the installed base for the Palm,
however, it’s only a matter of time before developers will create SOAP clients for it as well.
One of the best examples of a SOAP client for the Palm is kSOAP (http://ksoap.enhydra.org/
software/downloads/index.html). This Java implementation requires you to download
kXML (http://kxml.enhydra.org/software/downloads/index.html). Both products are in
beta now and you might find it difficult to create robust applications using them. However,
both will work with the Palm (at least the Palm VIIx I used for this book).
Another potential SOAP toolkit for the Palm is Bubbles (http://www.soaprpc.com/
software/bubbles/). Unfortunately, the site doesn’t contain downloadable code as of this
writing, but you might find some code by the time you read this. Like kSOAP, Bubbles
 Updating the Complex Type Example 329

relies on another product, KVM (http://java.sun.com/products/cldc/), to provide

low-level services. You’ll need both products to use this solution.

Updating the Complex Type Example

The Complex Type Example first appeared in Chapter 8. Remember that this is the example
that relies on 4S4C for the server-side component and pocketSOAP for the desktop. The
main reason for using third-party products is that the Microsoft SOAP Toolkit doesn’t pro-
vide the required support. In this case, we’re defining a complex string type that resembles a
Pascal string in that it has both a string part and a length part.
The following sections provide you with a detailed look at the client code and a summary of
some differences you should note between desktop and PDA code. We’ll also discuss some
important issues during application testing, such as sources of connectivity problems when
working with a PDA. This section contains some of the surprises developers experience
when they try SOAP on a PDA for the first time.
Ch

Creating the Client Code 10

You won’t need to change the server-side component for this example, so be sure to look
at the source code for it in Listing 8.2 of Chapter 8. Listing 10.1 shows the client code for
this example. Note that we’re using a Web page instead of the desktop application used in
Chapter 8. It’s interesting to note that you can move this application directly to a desktop
machine and run it within Internet Explorer without change.

Listing 10.1 Complex Type Client Code for PDA

<HTML>
<HEAD>
<TITLE>Simple PDA SOAP Example</TITLE>

<SCRIPT LANGUAGE=”JScript”>
function Test(String, Length)
{
var SOAPEnv; // SOAP envelope
var Transport; // SOAP transport
var Param; // Parameter list
var SOAPParam; // SOAP method call parameters.
var RecData; // Received data holder

// Create the envelope.

SOAPEnv = new ActiveXObject(“pocketSOAP.Envelope”);
SOAPEnv.MethodName = “GetComplexString”;
SOAPEnv.URI = “http://winserver/soapexamples/ComplexType/”;

// Create a parameter to place within the envelope.

Param = SOAPEnv.CreateParameter(“InString”, “”, “SOAPStruct”);

// Initialize the parameter.

SOAPParam = new ActiveXObject(“pocketSOAP.Param”);
SOAPParam.Init(“StringLength”, Length, “”);
330 Chapter 10 Working with PDAs

Listing 10.1 Continued

Param.Parameters.Append(SOAPParam);
SOAPParam = new ActiveXObject(“pocketSOAP.Param”);
SOAPParam.Init(“StringData”, String, “”);
Param.Parameters.Append(SOAPParam);

// Send the request and receive the data.

Transport = new ActiveXObject(“pocketSOAP.HTTPTransport”);
Transport.Send(“http://winserver/soapexamples/ComplexType/soap.asp”,
SOAPEnv.Serialize());
RecData = Transport.Receive();
SOAPEnv.Parse(RecData);

// Display the result.

RecData = SOAPEnv.Parameters.Item(0);
window.document.SampleForm1.Results.value = RecData.Value;
}
</SCRIPT>

</HEAD>

<BODY>
<!Add a heading->
<CENTER><H2>Get Complex String</H2></CENTER>

<!Use a form to test out the VBScript.->

<FORM NAME=”SampleForm1”>

<!Create a text entry control.->

Input a string:
<INPUT TYPE=text NAME=String VALUE=”Hello” SIZE=20><P>
Input a length:
<INPUT TYPE=text NAME=Length VALUE=”10” SIZE=20><P>

<!Define a place to put the results.->

Result values: <BR>
<TEXTAREA ROWS=3 COLS=40 NAME=”Results”>
</TEXTAREA><P>

<!-Create the two required buttons.->

<INPUT TYPE=button
NAME=cmdGetData
VALUE=”Get Data”
ONCLICK=”Test(String.value, Length.value)”>

</FORM>
</BODY>
</HTML>

Script is a case-sensitive language. This means that you have to look at capitalization
errors such as using Value in place of value. Unfortunately, the IDE for Visual Studio
does a poor job of pointing out such errors, so locating them can be time consuming.
Fortunately, Internet Explorer and other browsers will at least provide a line number
as part of an error message so you know which line of code to look at.
 Updating the Complex Type Example 331

The source code for this example follows the same five-step process as the desktop client.
First, you create an envelope. Second, you place data within the envelope. Third, you initial-
ize the data. Fourth, you send and receive the data. Finally, you display the data onscreen.
Notice that the HTML code for this example looks the same as HTML that you use in
any browser. Many developers now rely on the eXtensible HyperText Markup Language
(XHTML) (http://www.w3.org/TR/xhtml1/) to work with PDAs. Using XHTML has the
advantage of keeping the screens readable and small enough for a PDA. The disadvantage is
that XHTML often makes desktop displays unusable, or at least unfriendly. Make sure you
choose between HTML and XHTML at the outset of your development project based on
how you intend to use the Web pages later.

Differences in Implementation
Compare the code in Listing 10.1 with the client code in Listing 8.3 and you’ll see many
similarities. You need to recognize the limitations of a PDA, however. In this case, the code
will fail if you try to next calls in the same way that they appear in Chapter 8. For example,
Ch
in Chapter 8 you’ll see this nested call.
SOAPEnv.parse Transport.Receive 10
We need to replace that call with an alternative in this chapter. Note that the next call now
relies on an intermediate variable and requires two lines of code as shown here.
RecData = Transport.Receive();
SOAPEnv.Parse(RecData);

Of course, the use of JScript makes a difference as well. The HTML code looks similar to
C, while we used Visual Basic for the desktop example. These differences can cause prob-
lems if you don’t know what to look for. In most cases, a PDA will require extra steps to
accomplish the same task that a desktop machine can. This makes sense considering the
resources on a PDA are limited.

Testing the Application

Figure 10.1 shows the application in action. If you account for differences between a desk-
top and browser application, the two applications look about the same. They work the same
as well. Of course, the Web client design works with the smaller screen of a PDA. We’ll
discuss display size issues in the “Addressing PDA Display Issues” section of the chapter.
Now that you have two versions of the same application, you can perform tests to see how
application performance differs on a PDA when compared to desktop. You’ll find that the
PDA executes slower for the obvious reasons. Not only is the processor slower, but the
PDA is using a Web-based application. However, you might be surprised to see the small
performance degradation the application experiences. Depending on the complexity of
your application, you might find that users will gain access to data on the road without
losing much performance.
332 Chapter 10 Working with PDAs

Figure 10.1
The PDA version of
the application looks
and acts the same as
the desktop version.

One issue that will crop up is a problem with HTTP timeouts. An application that runs fine
on a desktop might suddenly experience errors on a PDA. A combination of factors contribute
to this problem, including
■ PDA processing speed
■ Connection speed
■ Web-based application performance issues
■ Server biases for processing local data first
■ PDA connectivity issues

In most cases, you can combat this problem by setting the timeout value of the server
higher. However, this won’t fix the problem all of the time. You might find that the PDA
has a terrible connection to the Internet or LAN. The user might not understand how
PDA connectivity works. In some cases, the server’s load is too high and it won’t respond
quickly no matter what you do. Unlike the server that issues Web pages on a daily basis, a
server that services SOAP applications will require better than average response time and
processing power.

Updating the Computer Name Example

The Computer Name example first appeared in Chapter 9, “Moving to Web-Based
Applications.” This is the two component example that relies on a server-side component
to perform the work and a parsing component to provide a SOAP-friendly interface. You’ll
also remember that this example plays on one of the advantages of the Microsoft SOAP
Toolkit—the ability to handle complex redirections. The 4S4C product fails when you
attempt to use it with this component combination.
As with the previous example, the code for the two components remains unchanged in this
example. You’ll find the code for the server-side component in Listing 9.1. The processing
component code appears in Listing 9.2. Make sure you test the application thoroughly
using the thick (Listing 9.3) and thin clients (Listing 9.4) before you move on to the code
in this chapter.
 Updating the Computer Name Example 333

The following sections provide a detailed look at the client code, discuss some message
traffic issues, and tell you about some important differences between a 4S4C and a
Microsoft SOAP Toolkit server implementation. We’ll also discuss application testing. In
this case, the example uses multiple components, which can lead to interesting debugging
problems when working with a PDA.

Creating the Code

The client code for this section is more complex than the previous example., First, it requires
two service buttons instead of one. Second, the Microsoft SOAP Toolkit also requires slightly
different input than 4S4C, so there are differences in the message formatting code as well.
Listing 10.2 shows the client source code for this example.

Listing 10.2 Computer Name Client Code for PDA

<HTML>
<HEAD>
<TITLE>CompName JScript Example</TITLE> Ch

<SCRIPT LANGUAGE=”JScript”> 10
function cmdGetSingleName_Click()
{
var SOAPEnv; // SOAP envelope
var Transport; // SOAP transport
var Param; // Parameter list
var SOAPParam; // SOAP method call parameters.
var RecData; // Received data holder

// Create the envelope.

SOAPEnv = new ActiveXObject(“pocketSOAP.Envelope”);
SOAPEnv.MethodName = “GetCompName”;
SOAPEnv.URI = “http://tempuri.org/message/”;

// Create a parameter to place within the envelope.

Param = SOAPEnv.CreateParameter(“NameType”,
window.document.SampleForm1.comboName.value, “”);

// Send the request and receive the data.

Transport = new ActiveXObject(“pocketSOAP.HTTPTransport”);
Transport.SOAPAction = “http://tempuri.org/action/NameValuesProc.GetCompName”
Transport.Send(“http://WinServer/soapexamples/ComputerName/CompNameProc.WSDL”,
SOAPEnv.Serialize());
RecData = Transport.Receive();
SOAPEnv.Parse(RecData);

// Display the result.

RecData = SOAPEnv.Parameters.Item(0);
window.document.SampleForm1.Results.value = RecData.Value;
}

function cmdGetAllNames_Click()
{
334 Chapter 10 Working with PDAs

Listing 10.2 Continued

var SOAPEnv; // SOAP envelope
var Transport; // SOAP transport
var Param; // Parameter list
var SOAPParam; // SOAP method call parameters.
var RecData; // Received data holder

// Create the envelope.

SOAPEnv = new ActiveXObject(“pocketSOAP.Envelope”);
SOAPEnv.MethodName = “GetAllNames”;
SOAPEnv.URI = “http://tempuri.org/message/”;

// Create a parameter to place within the envelope.

//Param = SOAPEnv.CreateParameter(“NameType”,
window.document.SampleForm1.comboName.value, “”);

// Send the request and receive the data.

Transport = new ActiveXObject(“pocketSOAP.HTTPTransport”);
Transport.SOAPAction = “http://tempuri.org/action/NameValuesProc.GetAllNames”
Transport.Send(“http://WinServer/soapexamples/ComputerName/CompNameProc.WSDL”,
SOAPEnv.Serialize());
RecData = Transport.Receive();
SOAPEnv.Parse(RecData);

// Display the result.

RecData = SOAPEnv.Parameters.Item(0);
window.document.SampleForm1.Results.value = RecData.Value;
}
</SCRIPT>

</HEAD>

<BODY>
<!Add a heading->
<CENTER><H1>Computer Name Component Test</H1></CENTER>

<!Use a form to test out the VBScript.->

<FORM NAME=”SampleForm1”>

<!Create a text entry control.->

Select a single computer name if needed:<BR>

<select name=comboName>
<option value=0>ComputerNameNetBIOS
<option value=1>ComputerNameDnsHostname
<option value=2>ComputerNameDnsDomain
<option value=3>ComputerNameDnsFullyQualified
<option value=4>ComputerNamePhysicalNetBIOS
<option value=5>ComputerNamePhysicalDnsHostname
<option value=6>ComputerNamePhysicalDnsDomain
<option value=7>ComputerNamePhysicalDnsFullyQualified
<option value=8>ComputerNameMax
</select><P><P>

<!Define a place to put the results.->

Result values: <BR>
<TEXTAREA ROWS=11 COLS=60 NAME=”Results”>
</TEXTAREA><P><P>
 Updating the Computer Name Example 335

Listing 10.2 Continued

<!-Create the two required buttons.->
<INPUT TYPE=button
NAME=cmdGetSingleName
VALUE=”Get Single Name”
ONCLICK=cmdGetSingleName_Click()>
<INPUT TYPE=button
NAME=cmdGetAllNames
VALUE=”Get All Names”
ONCLICK=cmdGetAllNames_Click()>

</FORM>
</BODY>
</HTML>

As with the previous example, this example uses a five-step process for each function. The
HTML code is harder to optimize for a small screen, in this case, because of the amount of
data we need to display. We’ll see later how you can overcome these problems without mod-
Ch
ifying the server-side component.
10
A Look at the Message Traffic
The Trace Utility (MSSoapT) allows you to monitor message traffic between the client and
server. Generally, you’ll use this tool to debug your application. However, you can also use
it to analyze the way clients and servers interact. Learning how to see major differences
between two implementations quickly can save you major debugging time.
Figure 10.2 shows the message traffic for the Thin Client example in Chapter 9. Remember
that this example relies on the Microsoft SOAP Toolkit and VBScript. Notice the message
includes only a few namespace entries. In addition, the message doesn’t define data types.
We’ve discussed this issue before with the Microsoft SOAP Toolkit.

Figure 10.2
An example of the
message traffic for
the Microsoft
SOAP Toolkit.
336 Chapter 10 Working with PDAs

Now let’s look at the pocketSOAP version of the message traffic in Figure 10.3. Notice
that this message contains far more namespaces and that the message defines the data
types. The pocketSOAP message provides more information and works with more
servers. One area where pocketSOAP might present problems is the XML tag at the
beginning of the message. Notice that this tag is missing and the server won’t know
which version of XML the client needs.

Figure 10.3
An example of the
message traffic for
pocketSOAP.

Note that although the client message has changed, the server response hasn’t. That’s because
the server-side component is unchanged. You’d need to change the server-side component
code to allow interoperability with other platforms that require it, such as Apache.
Figures 10.2 and 10.3 demonstrate something else. The basic format of the message is
the same. They have differences, but the two clients present the information in the same
way. The client needs to speak the same language as the server. This means changing
some message elements within your code to match the server’s implementation of SOAP.
I hope that these changes will become fewer as vendors work on the SOAP specification.
The importance of looking at message flow now is being able to create clients that will work
with any server quickly. This means analyzing differences between the PDA client and a
desktop client that already works. Because your choices of PDA SOAP toolkits are limited,
you’ll want to learn to recognize problem areas in messages.
 Updating the Computer Name Example 337

Server Differences Revisited

Listing 10.1 and Listing 10.2 have some differences that are obviously due to distinctions
in application. However, you might not have noticed the nuances of difference between
the two messages that are server related. For example, look at the Complex Type output
message in Figure 10.4. Notice that some of the namespace URLs differ from those in
Figures 10.2 and 10.3.

Figure 10.4
Server differences
become more appar-
ent when you analyze
message traffic.

Ch

10

When working with the Microsoft SOAP Toolkit, you’ll always set the SOAP envelope URI
to http://tempuri.org/message/. The 4S4C server always requires the base URL for an
application as part of the envelope. Although both servers rely on WSDL to define the
service, they do so using different techniques.
Another difference you should notice is that the Microsoft SOAP Toolkit requires a
SOAPAction entry. You create this entry as part of the transport, not the envelope as you
might expect. If you look at Figures 10.2 and 10.3, you’ll notice that the SOAPAction entry
doesn’t even show up. We looked at another utility, tcpTrace, that does show this particular
entry. Although the Trace Utility provides a neater, easier-to-read display, you’ll need
tcpTrace to show differences of this sort. If you don’t include the SOAPActor, the server
will complain and the message will remain unprocessed.
338 Chapter 10 Working with PDAs

You might want to check the source code for other differences. For example, the transport
URL is precisely the same when using pocketSOAP as it is when using the Microsoft client.
However, the transport URL for pocketSOAP when using 4S4C is http://localhost/
soapexamples/ComplexType/soap.asp in this example. Notice that we didn’t include the
“?WSDL” part that normally appears in the Microsoft SOAP client code. Again, this is
another example of server differences; it also shows how clients interact with their server.

Testing the Application

The Computer Name example doesn’t work quite the same on a PDA as it does on the
desktop. The reason is simple—a PDA lacks the screen real estate to display this application
fully. Part of the testing process is to ensure the screen remains close to the original (to keep
training costs down) while ensuring readability. Figure 10.5 shows the output from this
example if you click Get Single Name.

Figure 10.5
Screen real estate
makes a difference in
application appear-
ance in some PDA
applications.

The application does work the same from a functionality perspective. You still use a drop-
down list box to select a computer name type. The buttons still produce the same results as
before. Maintaining this level of compatibility between a desktop and PDA application is
essential to the success of the application. Any application that requires the user to relearn a
business process is doomed to failure, no matter how nice the display looks.
Fortunately, this example is reasonably small. Your testing process must include resource
usage and other factors that affect PDA application performance. Complex database applica-
tions won’t run very well on a PDA because they require too many resources. Of course, you
can always get around the problem by using more server-side processing and presenting the
PDA with semi-finished, rather than raw results.

Addressing PDA Display Issues

Most developers realize that working with a PDA is going to be a challenge before they
begin their first project. The problem is that the challenge usually turns out larger than
expected. Small displays with limited color capability are something that most developers
consigned to the past. The old techniques that developers used to conserve screen space are
suddenly appearing again.
 Addressing PDA Display Issues 339

The following sections discuss several important display issues when working with a PDA.
This section won’t provide an in-depth treatise on the subject, but you’ll walk away with
some fresh ideas for your next SOAP application.

Screen Size
Many users have 17” or 19” monitors capable of a minimum of 1280 × 1024 resolution
today. Developers have taken advantage of the screen real estate to create better applica-
tions that display more data at one time. Even Microsoft uses higher resolutions as a
baseline for applications—many of their application screens won’t fit on a 800 × 600
display anymore.
Everything you want to do with your PDA has to fit within 320 × 200 pixels. That’s a lot
smaller than the typical computer screen. In addition, some PDAs use black-and-white
displays in place of color, so you can’t even use some of the modern tricks to make the
display look nicer. In short, PDA screens tend to look a bit plain, and developers normally
find themselves feverishly cutting their application screens down to size.
Ch
No matter what you do, it’s impossible to fit 1280 × 1024 worth of application screen in a
320 × 200 space in many cases. When this happens, you’ll find that you need to make some
10
compromises in the display. For example, even the computer name example in this chapter
ran into problems displaying all of the information that it could provide. Figure 10.6 shows
the results of some cutting that I performed to make the application data fit and still look
reasonably nice.

Figure 10.6
PDA displays are
challenging to work
with because they’re
so small.

In this case, I indented the second line of each data entry to allow enough space for long
entries. Notice that the data is still readable and the user won’t have to guess about the
formatting. Of course, this is still a less than perfect solution because the data does appear
on two lines.
It’s important to keep every application element on a single screen if possible. Figure 10.6
does this by sacrificing application data display space. The user can scroll through the data
in the Result values field without moving other screen elements around. The stylus provided
with a PDA doesn’t lend itself to mouse-like movement.
340 Chapter 10 Working with PDAs

Make sure you consider XHTML for complex applications with many elements. XHTML
allows you to display your application in segments with relative ease. Other options include
using the Handheld Device Markup Language (HDML) (http://www.w3.org/TR/NOTE-
Submission-HDML-spec.html) or Wireless Markup Language (WML) (http://www.
oasis-open.org/cover/wap-wml.html). Both of these technologies use the concept of cards
and decks to break up information into easily managed pieces. Of course, the PDA you use
has to provide support for these standards before you can use the tags within a document.
As with XHTML, using either HDML or WML will prevent your
page from appearing properly on a desktop machine.

Using Color
Developers have gotten used to seeing colors on their applications. Color dresses up a drab
display and makes the application more fun to use. In addition, using color presents cues to
the user. For example, many users associate green with a good condition and red with some-
thing bad. In short, most applications today rely heavily on color with good reason.
Depending on the PDA you use, you might not have color at all. For example, many Palm
models present the world in shades of gray. Even if a PDA does provide color support akin
to the Pocket PC, the developer still has to use color carefully.
The problem for PDA users is that the screen is already small. If they get into an area with
bright sunlight, seeing the screen might become impossible, especially if it’s filled with
colors that don’t work well in such an environment. (Vendors are making advances with
regard to bright sunlight. For example, the Ipaq H3650 provides good display support even
in bright sunlight.) Notice that the PDA screenshots in this chapter are mainly black and
white. The actual screens contain some color for the icons, but that’s about it. Because these
applications don’t need color to present the information they can provide, it’s possible to
rely on a black-and-white image.
Using color to display icons or to convey a message is still a good idea, even in the world of
the PDA. For example, a red icon could signal danger or tell the user to wait without using
up screen real estate for words. Of course, you need to explain the meaning of the color
changes within a manual or help file (preferably both).

Pointer Pointers
Most PDA users rely on a pointer to do all of their work. Sure, a few PDAs do offer a key-
board and mouse as separate items, but most of these offerings are bulky and difficult to use.
Pointer use is one of the reasons that you want to keep your application on one screen, or
use multiple screens when necessary. Scrolling on a PDA screen is less than intuitive and
requires some level of skill to master.
SOAP applications that you move from the desktop to the PDA will require some modifica-
tion for screen size in many cases. While you’re working on the screen, it might be a good
time to add some pointer-friendly features as well. For example, try to make as many tasks
as possible accessible with a single pointer touch. Users should be able to point to what they
want and allow the PDA to complete it for them.
 Understanding PDA Security Issues 341

You can also build intelligence into the application. A SOAP application normally has a direct
connection to the server. You can use some of the server’s processing power to make things
easier on the user. Most PDAs already include predictive logic as part of their setup. For
example, as you write something, the PDA tries to guess the entire word. When it guesses
the correct word, you can click on it and save some writing time. The same principle works
for other activities as well. For example, a SOAP application could automatically display a
data screen that the user needs most often, rather than force the user to dig through several
screens to find it.
Pointer-friendly programs also make tasks yes or no propositions. Again, this allows the
user to accomplish the task with a single click, rather than write something down. The
point is to make the PDA as efficient as possible so the user doesn’t get frustrated trying
to do something easy.

Beyond PDAs to Telephones

Ch
You won’t find telephones that run SOAP applications today. In fact, I was surprised to find
that some vendors are already creating embedded SOAP applications. (We talked about 10
these applications in Chapter 9.) Embedded applications reside within toasters, your car, and
even your television set. As SOAP moves from the desktop to the PDA, it’s only logical that
it will move to the telephone as well.
SOAP is a good technology for telephones and other small devices because it provides
s simple protocol. Telephones are all about communication. As more people start carrying
cellular telephones, the need for connectivity applications will increase. This is an exciting
time for SOAP developers because of the many different devices with which to work.
The interesting thing about telephones is that we already use them for so many business
and personal purposes. For example, it’s possible to check your bank statement or send
correspondence to your credit card company using the telephone. All you need to know
how to do, in most cases, is push buttons.
Of course, speaking into the telephone or entering some numbers for an account is hardly the
same thing as using an application on a desktop computer. SOAP will enable two-way commu-
nications in a way that users haven’t seen before. For example, an employee could enter order
information using a telephone (assuming the orders are small and relatively simple).
Will telephone SOAP applications appear overnight? Probably not. Will we use the tele-
phone to accomplish the same things we do on the PC? Again, probably not. However, the
telephone does present some interesting opportunities for SOAP developers in the future.

Understanding PDA Security Issues

Many network administrators view the PDA with more than a little suspicion—for good
reason. The media has painted the PDA as a device that is so open that anyone can access it at
any time. Smart users keep all of their sensitive data on the company computer and just place
342 Chapter 10 Working with PDAs

lists of tasks on their PDA. Of course, such a view defeats the entire purpose of having a PDA
in the first place. A PDA should be an extension of your workplace, not a hindrance to avoid.
Many of the security issues surrounding PDAs today are a perception that they’re all wire-
less devices. Many PDAs use network or modem connections, not wireless connections. The
PDAs that do provide wireless access tend to err on the side of safety whenever possible.
Vendors realize that wireless access is both a blessing and curse for many companies.
However, the wireless issue probably isn’t much of an issue for the SOAP developer. Count
on many of your users to rely on modem or direct network interface card (NIC) connec-
tions. For example, the Pocket PC provides a slot that will accommodate an Ethernet card
that provides a direct network connection.
One of the biggest security issues for all PDA users is data storage. There are actually two
threats to consider. The first is that someone could access sensitive data if he stole your
PDA (or at least borrowed it for a while). Wireless or not, the person who has the PDA also
has access to the data it contains. The second threat is data loss. Many PDAs lose all of their
information after the main battery is exhausted. (Some PDAs provide a backup battery that
retains the contents of memory until the user replaces or recharges the main battery.) Unless
the user backs up on a regular basis while on the road, the data is lost.
SOAP can come to the rescue, in this case, by allowing the user to store all data on a central
server in a secure location. The application remains on the client machine, but the data
remains secure at the main office. Of course, some level of flexibility has to be built into the
security plan so that users can operate their PDA in a disconnected mode. PDA vendors will
likely add some form of biometric protection to their PDAs in the future. For example, the
user might have to press his thumb in a certain area in the back when starting the PDA so
the PDA can check his identity.
Some security issues involve limitations in the PDA itself. For example, when working with
a desktop application, you can ask the user to enter his name and password. The application
asks the server for verification before it requests data access. The pointer a PDA user must
use to enter data hampers the security effort. It’s possible that he’ll end up entering the pass-
word more than the three times that good applications normally allow. The next thing you’ll
hear is a frustrated user on the telephone asking why he can’t access his application. Security
is required, but any attempt to implement security that causes user frustration is almost
certain to fail.
Another PDA-specific security issue to consider is one of resources. Heavy security con-
sumes many resources. A desktop user notices some slowing of his computer when he’s
working with a secure application. The slowing is due to the work of the security protocol
in the background. With the limited resources of a PDA, you probably can’t implement a
level of heavy security because the application would run too slowly. The PDA already runs
SOAP applications slower because it uses a browser-based application over a relatively slow
connection. Adding to this problem is one way to lose user confidence in your application.
 Troubleshooting 343

PDAs do suffer from every other security problem that SOAP applications normally experi-
ence. A lack of built-in security will allow prying eyes to see the data that the user transfers to
the home office (at least if someone is actually looking). Unfortunately, the list of available
security add-ons for PDAs is extremely limited. Theoretically, you can use technologies such
as SSL with a PDA. We’ve already discussed the limitations of some of these technologies in
previous chapters, so I won’t go into them again here. Needless to say, using a PDA involves
some security risk that you’ll need to overcome.

Troubleshooting
This chapter has shown you how to work with PDA applications under SOAP. PDAs pre-
sent some interesting challenges no matter which programming language you use to develop
a client and no matter how you transfer data. Fortunately, unlike many technologies on the
market, SOAP allows you to preserve the server side of your development investment. If you
use a PDA such as the Pocket PC, you can also reuse some of your client code. In short,
SOAP solves significant PDA development problems. The following sections will examine
some of the questions developers ask about using SOAP for PDA development and hope-
fully provide the answers you need to add SOAP to your PDA application toolkit. Always
feel free to contact me at [email protected] if you have additional questions.

Why Does the Example Seem to Run, and Then Display Nothing Onscreen?
The number one problem for PDAs running applications from a Web server is security.
The PDA will often run into an access problem and simply ignore it. You won’t see an
error message in many cases. All that will happen is that the PDA will perform whatever
local tasks it can perform and then stop when the error happens.
You’ll also see this problem if you use a scripting language the PDA doesn’t understand.
Again, the PDA will try to accomplish all that it can locally, and then it’ll simply stop.
Remember that Windows CE doesn’t provide VBScript support, but it does provide JScript
support. Other PDA setups have similar limitations, and you need to check for them during
the project definition stage.
Some developers ignore the small screen that a PDA provides. You might not see a result
because it appears offscreen. Moving the scrollbar up and down might show an answer that
you couldn’t see before. Unfortunately, users will normally complain about the lack of data
long before they move the display. In short, make sure that your application fills a single
screen at a time so the user doesn’t have to look for the information.
In a very few cases, you’ll find that all of the client files need to appear on the PDA instead
of the Web server. For example, downloading a Web page so that the PDA can access the
SOAP scripts might not work. Test all of the machines your application targets to ensure
you won’t run into this type of problem.
344 Chapter 10 Working with PDAs

Why Doesn’t My Code Run Properly on All of My PDAs?

During the development process for the examples in this chapter and a few personal projects,
I found that even though a PDA runs the same operating system as another PDA, you need
to account for some differences. For example, PDAs use different processors, so you can’t
assume anything about low-level code. Consequently, it’s important to use high-level
languages if you plan to work with the code on more than one machine.
You’ll also find that PDAs require driver and other software updates, just as desktop
machines do from time-to-time. Make sure you download and install all patches for your
PDA before you begin development. In fact, you’ll want to perform this step before you
install the SOAP toolkit of your choice. This ensures that the installation routine for the
SOAP toolkit makes the right choices about your system.
Some SOAP implementations require support files. Make sure you install all of the required
support before you attempt to run an application on the PDA. For example, the Windows
CE machines in the example all required a copy of pocketSOAP before they would run the
example. The Palm systems required both the SOAP toolkit and the XML parser. Palm
systems require two separate installations, so leaving one part of the support mechanism out
of the picture is relatively easy.

How Do I Fix Messaging Problems with the Client?

Most messaging problems appear when you use more than one toolkit to accomplish a task.
The server might expect something other than the normal client output. However, you can
normally modify the data stream enough to accommodate the requirements of client or
server. The various tracing tools allow you to compare message formats and change the
standard code as needed to create a message the server will understand. Likewise, you can
normally modify the server output to accommodate client needs.
Another problem is live connection confusion. A PDA requires a live connection with the
server, not a synchronized connection. A live connection allows data exchange between client
and server, while a synchronized connection allows the host workstation to update the PDAs
files. In some cases, a synchronized connection might look live, but it isn’t. PDAs always require
some type of direct network connection to provide data connectivity, plus a driver that allows
network configuration. In short, don’t assume all wireless connections are live simply because
they tap into the network—many simply provide a better way to perform synchronization.
Always double-check a failed PDA connection using a desktop client that you know works.
PDA connections can fail in numerous ways that a desktop connection won’t. In at least a
few cases, you’ll find that the server has stopped responding and the desktop application
will provide an error message to this effect. PDAs often suppress error messages, making
debugging especially difficult.
 APPENDIX
A
SOAP Data Types and Data Type
Conversions
In this appendix
Data Types Overview 346
Complex Data Types 349
Differences in Implementation 351
Data Type Conversions 353
346 Appendix A SOAP Data Types and Data Type Conversions

Any data exchange technology worth its salt has to include the concept of data types. SOAP
is no exception. You’ll find that SOAP messages use a simple, but usable set of data types. Just
how usable these data types are depends on how the SOAP toolkit you use implements them.
As I write this, there are several different prevailing views of data typing with SOAP and
these views affect how SOAP toolkit vendors implement the specification. The SOAP speci-
fication doesn’t address the issue of data typing itself, but instead defers to several XML
specifications. You’ll find these XML specifications listed in the SOAP specification
(http://static.userland.com/xmlRpcCom/soap/SOAPv11.htm).
While you’re looking at the SOAP specification, it’s important to note two of the author
names on this document. The first is Don Box, the second Dave Winer. These two individ-
uals have contributed a great deal to the current specification. However, when you talk with
them online, you find that they have very different viewpoints on the current SOAP imple-
mentation, and where they would like to see it go in the future. (Many other authors con-
tributed to the SOAP specification, but Don Box and Dave Winer seem to be two of the
more vocal contributors.)

Don Box and Dave Winer are both prolific writers, along with being great computer sci-
entists. You can read Don’s current assessment of the state of SOAP at http://www.
develop.com/dbox/postsoap.html. Compare this to Dave’s view of the current
state of SOAP at http://www.xmlrpc.com/stories/storyReader$555, http://
www.xmlrpc.com/stories/storyReader$1387 and http://www.xmlrpc.com/
stories/storyReader$1433. Reading these four pieces gives you a good under-
standing of why things work the way they do today. On one hand, you have Don’s
enterprise view of SOAP—on the other hand, you have Dave’s developer view of SOAP.

So, what does this have to do with data types? A data type is more than a simple definition of
a string or a float. Knowing that data of a certain type exists isn’t enough to allow data
exchange. This appendix addresses the issue of data types in a fuller sense of the term. We’ll
discuss how SOAP applications exchange data and what you might need to do to ensure inter-
operability between SOAP implementations. Obviously, the whole issue of SOAP data types is
under
constant change. You’ll want to check the SOAP specification. Also check online list servers
such as the one found on the Developmentor Web site (http://discuss.develop.com/) and
Microsoft’s newsgroups (microsoft.public.msdn.soaptoolkit, microsoft.public.xml.soap, and
microsoft.public.xml.soapsdk) on the news://news.microsoft.com news server.

Data Types Overview

Many developers will rely on SOAP to transmit data in text format and perform any required
data translation at the endpoints. In many cases, this is the best way to ensure interoperabil-
ity, but it does entail extra work for the developer and affects the performance of the applica-
tion. SOAP supports a relatively short list of simple and complex data types. You’ll find the
 Data Types Overview 347

simple types defined at the SOAP encoding schema Web site (http://schemas.xmlsoap.
org/soap/encoding/). Figure A.1 shows some of what you’ll see at this Web site. Table A.1
describes each of these types. It also tells you how to use them within a message.

Figure A.1
You’ll reference the
SOAP encoding
schema as part of
most messages.

App

A
Table A.1 SOAP Simple Data Types
Data Type Usage
base64 Non-ASCII data encoded using the RFC2045 rules (http://www.faqs.org/rfcs/
rfc2045.html). This data type enables you to transfer data using the same techniques
as Multipurpose Internet Mail Extensions (MIME) without the line length limitations.
binary Non-ASCII data encoded using standard hexadecimal methods. You normally avoid this
data type because it doesn’t transfer well through firewalls.
boolean A value of true or false represented by a 1 for true and 0 for false. Boolean representa-
tions differ among programming languages, which means you may have to convert this
value to a valid language representation after the XML parser interprets it.
byte A data type derived from short where the maximum inclusive value is set to 127 and the
minimum inclusive value is set to -128.
date A Gregorian date representation as defined by ISO8601. The date uses a format of
<4-digit year>-<2-digit month>-<2-digit day> (that is, 2001-03-31 for March 31, 2001).
It may include an optional time zone value that helps in ordering dates.
decimal A decimal (base 10) representation of an integer used to increase the accuracy of some
types of calculation such as financial data. The sender must indicate the precision of the
decimal value (normally 18-digits) and the number of fractional digits within the number.
double An IEEE 64-bit precision floating-point number as defined by IEEE 754-1985 (http://
standards.ieee.org/reading/ieee/std_public/description/ busarch/754-
1985_desc.html).
348 Appendix A SOAP Data Types and Data Type Conversions

Table A.1 Continued

Data Type Usage
float An IEEE 32-bit precision floating point number as defined by IEEE 754-1985.
int A data type derived from long where the maximum inclusive value is set to
2,147,483,647 and the minimum inclusive value is set to -2,147,483,648.
integer A data type derived from decimal where the number of fractional digits is set to 0.
This differs from most language integer definitions and avoids some of the problems
associated with supporting multiple platforms. Note that integer is the type used to
define all other integer data types.
long A data type derived from integer where the maximum inclusive value is set
to 9,223,372,036,854,775,807 and the minimum inclusive value is set
to –9,223,372,036,854,775,808.
short A data type derived from int where the maximum inclusive value is set to 32,767 and
the minimum inclusive value is set to -32,768.
string The sequence of human readable characters normally associated with text.
time The representation of a single instant of time in a 24-hour day. The time data type
uses a format of hh:mm:ss.sss in accordance with ISO8601. This data type may
include an optional time zone value that helps in ordering dates.

Table A.1 doesn’t contain data types that you’d normally use in areas other than data trans-
fer elements. For example, it doesn’t include the uriReference type. The specification also
includes support for some derivations of simple data types such as unsignedLong and
negativeInteger. Some data types are actually subsets of the types listed in Table A.1 such as
month, day, and year. You can find out more about these other types by looking at the
encoding schema Web page. Of course, this Web page provides little in the way of explana-
tion—it only tells you about implementation. You can read the descriptions for all of the
data types that XML supports at http://www.w3.org/TR/xmlschema-2/. This specification
doesn’t contain descriptions for data types unique to SOAP.
It’s important to differentiate between the simple types that XML supports and those sup-
ported by SOAP. The SOAP specification is unclear on this issue, but the schema encoding
Web page will clear things up for you. You’ll find that SOAP supports most XML data
types. However, some simple data types receive support as derived types. For example, the
XML duration data type is timeDuration and recurringDuration under SOAP. Likewise,
SOAP doesn’t support the dateTime XML data type, but does support individual date
and time data types. SOAP supports a binary data type that doesn’t appear in the XML
specification—XML supports only hexBinary and base64Binary data types.
The SOAP specification also places enumerated values within the simple type category. You
use an enumeration to define a specific set of values for a user-defined data type based on
the simple data types found in Table A.1. Here’s an example of an enumeration.
<element name=”Direction” type=”tns:Direction”/>
<simpleType name=”Direction” base=”xsd:string”>
<enumeration value=”Forward”/>
 Complex Data Types 349

<enumeration value=”Backward”/>
<enumeration value=”Left”/>
<enumeration value=”Right”/>
</simpleType>

<Car>
<WhereTo>
<Direction>Forward</Direction>
<Speed>70</Speed>
</WhereTo>
</Car>

In this case, a car is going forward at a speed of 70. The enumeration limits the number of
valid values to forward, backward, left, and right. This means that the car couldn’t go up, for
example. The use of an enumeration helps document the SOAP message. It also provides a
form of validation for the message. However, you need to use enumerations with care
because they can also limit the flexibility and extensibility of a schema that you create.

Complex Data Types

Sometimes it helps to group data together in meaningful arrangements. Complex data types
contain arrangements of simple data types. However, the declaration of a complex type is
similar to a simple type as shown in Figure A.2. The key declarations tell you how SOAP
will use various keywords such as type. Note that the declaration of complex and simple
types is mutually exclusive as defined by the <selector> tag. SOAP supports two complex App
(or compound) data types: array and structure. A
The SOAP specification defines many ways to use the two complex data types that it
supports. However, before you can actually use complex data types, your SOAP toolkit
needs to support them. Support for complex data types is getting better, but some
SOAP toolkits don’t support them yet. Even if your SOAP toolkit supports complex data
types, there isn’t any guarantee that the implementation will fit in with the implementa-
tion used by other vendors. The use of complex data types represents one of the major
reasons for interoperability problems in applications right now. In short, use complex
data types with care. Exploit complex data types only if you plan to use the same SOAP
toolkit for the entire application.

Structures normally pass a single instance of related data from one point to another. For
example, you could view a single database record as a type of structure. Here’s a simple
schema of a structure under SOAP.
<element name=”Contact”>
<complex type>
<element name=”ContactName” type=”xsd:string”/>
<element name=”Address” type=”xsd:string”/>
<element name=”Telephone” type=”xsd:string”/>
</complex type>
</e:Contact>
350 Appendix A SOAP Data Types and Data Type Conversions

Figure A.2
It pays to look at the
various key declara-
tions because they
show how SOAP uses
specific elements.

This schema defines a complex user-defined data type named Contact that contains three
strings. Each string defines a particular element of the contact entry. Here’s this schema
in action.
<e:Contact>
<ContactName>George Smith</ContactName>
<Address>100 North Street</Address>
<Telephone>(303)555-1212</Telephone>
</e:Contact>

As you can see, the schema packages the data and makes it easier to validate. The package
arrives on the other end of the wire in a format that’s easy to parse into a database. While
the names within the structure are significant, the order of the elements isn’t. The SOAP
message can contain the structure elements in any order so long as the sender doesn’t violate
any of the structure schema rules.
Structures can use all of the normal XML extras such as identifiers and references. You create
an identifier using the id attribute within the SOAP message like this <e:Contact
id=”MyContact”>. The reference uses the familiar href like this, <AContact href=”#MyContact”>.

Arrays allow you to access individual elements of grouped data. Unlike many programming lan-
guages, SOAP doesn’t force you to use the same data type for all array elements. You can also
create ragged arrays with a little creative programming. Here’s an example of a simple array.
<element name=”anArray” type=”SOAP-ENC:Array”/>

<anArray SOAP-ENC:arrayType=”xsd:ur-type[3]”>
<myNumber xsi:type=”xsd:int>3</myNumber>
<myString xsi:type=”xsd:string>”Hello World”</myString>
<myBoolean xsi:type=”xsd:boolean>true</myBoolean>
</anArray>
 Differences in Implementation 351

The use of the ur-type allows you to define an array containing elements of mixed types. You
can also use a simple method to create an array of the same elements without having to
define each element type individually. Here’s an example of an array containing just one type.
<element name=”numberArray” type=”SOAP-ENC:Array”/>

<numberArray SOAP-ENC:arrayType=”xsd:int[3]”>
<myNumber>3</myNumber>
<myNumber>4</myNumber>
<myNumber>5</myNumber>
</numberArray>

You can create arrays of arrays or structures. Like structures, you can also include identifiers
and references in your array as long as you meet any required schema restrictions. In short,
arrays allow you to move data in a way that allows easy access of individual data elements.
You can use arrays in SOAP as you would in any programming language, with the appropri-
ate caveats, of course. While many SOAP toolkits support some form of array, some don’t
support anything more than single dimension arrays now. In addition, you’ll find that sup-
port for ragged arrays is lacking. Of course, you can always go the hand-coding route if you
want to implement a solution today that the usual assortment of toolkits doesn’t support.

Differences in Implementation
Now that you understand the data types, it’s time to consider how vendors implement these
App
data types. The technique used implement the data type determines where the type defini-
tions appear within the SOAP message. There are two common methods in use by most A
SOAP toolkit vendors now.
■ Define the data types as part of the elements in the SOAP message.
■ Define the data types as part of an external file that the listener or XML parser uses to
decode the SOAP message.

Let’s begin by looking at the first method for defining data types. Many developers feel that
using external files such as WSDL unnecessarily complicates what should be a simple proto-
col for exchanging data. Here’s a simple SOAP message that contains data type information
as part of the message.
<?xml version=”1.0” encoding=”UTF-8” standalone=”no”?>
<SOAP-ENV:Envelope
SOAP-ENV:encodingStyle=”http://schemas.xmlsoap.org/soap/encoding/”
xmlns:SOAP-ENC=”http://schemas.xmlsoap.org/soap/encoding/”
xmlns:SOAP-ENV=”http://schemas.xmlsoap.org/soap/envelope/”
xmlns:xsd=”http://www.w3.org/1999/XMLSchema”
xmlns:xsi=”http://www.w3.org/1999/XMLSchema-instance”>
<SOAP-ENV:Body>
<MyObj:GetPerson xmlns:MyObj=”http://www.mycompany.com/myobjects/”>
<LastName xsi:type=”xsd:string”>Mueller</LastName>
<FirstName xsi:type=”xsd:string”>John</FirstName>
</MyObj:GetPerson>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope
352 Appendix A SOAP Data Types and Data Type Conversions

Notice that two new namespaces are used in this example, xsi and xsd. Some forms of inline
type specification dispense with the namespaces and simply use type=”variable type” by itself.
The SOAP specification shows other methods of formatting the information in sections 5.2 and
5.3. You can also find other valid formats at http://www.lwprotocols.org/xmlrpc2soap.html. In
this case, the author is showing how to convert XML-RPC documents to SOAP format, but
the instructions for creating a data representation will help anyone.
Inline data typing allows the recipient to parse the data without using an external file. The
advantage of this technique is the SOAP message is easier to understand, making it easier to
debug an application. Using this technique also reduces ambiguity. Inline data typing allows
the receiving party to use the data without having to make any value decisions.
One of the problems using inline data typing is that it increases the size of the SOAP message.
This is a problem for most organizations because bandwidth is still somewhat precious, even
in today’s connected world. External file advocates stress that their method reduces the size of
messages, making SOAP more efficient. Using an external file such as a WSDL file also allows
the recipient to validate data against a known standard, potentially reducing errors.
The most popular external file methodology right now is WSDL. Here’s a short excerpt
from a WSDL file showing the data typing in action.
<message name=”AddItClass.DoAdd”>
<part name=”Add1” type=”xsd:short” />
<part name=”Add2” type=”xsd:short” />
</message>
<message name=”AddItClass.DoAddResponse”>
<part name=”Result” type=”xsd:short” />
<part name=”Add1” type=”xsd:short” />
<part name=”Add2” type=”xsd:short” />
</message>

As you can see, a WSDL file defines a schema for a given component. It provides standard-
ized variable names and assigns them a type. The ASP listener uses the data types in this file
to validate the input data, making it less likely that an application will use bad data for a
calculation or other processing need.
We’ve discussed many other interoperability problems as the book has progressed. Vendors
will fix many of those problems because they’re relatively easy to resolve. However, the issue
of using inline data typing or relying on an external problem isn’t easy to resolve. The stan-
dards committees spend a great deal of time resolving this issue and some vendors may
choose not to comply in the end.
The bottom line is that you need to choose a SOAP toolkit that provides the full range of
data type support that you require, using the technique that you’re most comfortable using.
As your application increases in size, you’ll have to face the probability that you’re going to
run into data related interoperability problems. These problems won’t vanish into the
night—they aren’t a phase of the moon problem.
 Data Type Conversions 353

Data Type Conversions

As previously mentioned, strings are the easiest way to transfer data from one point to another
when interoperability is the biggest concern. You shouldn’t have to resort to strings in most
cases, but it’s important to know that they present the least problematic way to move data. The
first data type conversion that you need to consider is moving everything to strings. Numeric
data is easy to convert, but some data types are harder. For example, you’ll need to create a
representation for Boolean values that the recipient won’t confuse with a standard string value.
Other data conversion issues aren’t as terrifying as converting everything to a string, but
they still represent problematic situations. For example, SOAP doesn’t provide any means to
move an object from one place to another. This makes sense since DCOM and CORBA
don’t move objects either. These protocols make it appear that you’re moving an object by
creating an object of the same type on the remote system using the same data as the object
on the client. The difference is that DCOM and CORBA hide the implementation details
from you. Given SOAP’s mission, it may not even be possible to move objects. A Windows
object is meaningless on a Linux server. In many cases, data conversion means changing the
volatile data contained in an object into a more permanent form.
You’ll notice that the XML specification provides very specific ranges for the various integer
types. Of course, you still have to consider the requirements of the local system. A short on
an Intel machine may not match the short used by SOAP. In most cases, it pays to create a
set of conversion routines that you call from your programming language of choice when App
performing any integer size conversions. Fortunately, Microsoft’s .NET Framework should
have this support built in by the time you read this. Users of older compilers will still need A
to create the conversion routines by hand.
It usually pays to follow the least common denominator approach with SOAP. In other
words, don’t use an integer when a string will do. Since strings are the easiest data type to
transport, make sure you use them as often as possible and convert when required. The fol-
lowing list contains the hierarchy I normally follow in determining which data type to use.
It contains all the data types listed in Table A.1 from least to most difficult.
■ string
■ integer
■ byte, int, long, short
■ decimal
■ boolean
■ double, float
■ date, time
■ base64
■ binary
354 Appendix A SOAP Data Types and Data Type Conversions

Data-conversion is a serious issue for developers who want to convert an application to

SOAP. In some cases, you’ll find that you need to restructure an application completely just
to comply with twisted data conversion requirements.
 APPENDIX
B
Microsoft BizTalk and SOAP
In this appendix
What Is BizTalk? 357
How BizTalk and SOAP Work Together 358
An Overview of Useful BizTalk Utilities for SOAP 359
BizTalk Fixes a Few SOAP Problems 368
Is BizTalk the SOAP Add-On for Your Company? 369
356 Appendix B Microsoft BizTalk and SOAP

Many developers wonder if they should forgo learning SOAP using toolkits and proceed
directly to one of the business-to-business (B2B) productivity suites such as BizTalk. Most
B2B suites today include some form of online connectivity. BizTalk happens to use several
communication techniques, including SOAP, to support online connectivity. This Appendix
examines the relationship between BizTalk and SOAP.
BizTalk is a huge product. An author could write an entire book on the product and not
even begin to scratch the surface of what this product could do for a business. With that in
mind, this Appendix will focus on the SOAP elements of BizTalk. I won’t assume that you’re
familiar with this product, but I’m not going to provide you with anything approximating a
complete overview of the product either.

BizTalk is an incredibly complex product with an equally complex set of installation

requirements. Make sure you read all of the installation material before you begin the
installation. You’ll likely find that you need to perform some configuration of your
server and download service packs for several server products, such as SQL Server,
before you begin the installation.

However, this Appendix is going to be useful for several reasons. For one thing, we’ll look at
the SOAP end of BizTalk in some detail, which will allow you to make better decisions
regarding the route you want to take with SOAP. If your business is ready to make a big
move online and you plan to spend most of that time working with other businesses, a B2B
solution might be better than the SOAP toolkits we’ve discussed throughout the book.
Even if you don’t use BizTalk for your main business application immediately, you can still
use the tools it provides to design your application. One of the tools will allow you to gener-
ate an XML Data Reduced (XDR) file of the type required by the SOAP Messaging Object
(SMO) framework. BizTalk also ships with a number of design tools that you can use alone
to improve your applications.
In short, getting the evaluation copy of BizTalk will be a smart move even if you don’t use the
product immediately because you can use the tools the evaluation product provides for experi-
mentation purposes. This Appendix will help you see the usefulness of BizTalk for any SOAP
developer. However, you’ll find that other vendors are also creating B2B solutions and will
probably want to look at them as well before you make a B2B product-purchasing decision.

Microsoft provides a full-featured 120-day evaluation copy of BizTalk that you can down-
load from their site (http://www.microsoft.com/biztalk/evaluation/trial/
default.asp). The 120-day evaluation period normally provides enough time to create
some simple test applications and determine if you want to use BizTalk or rely on other
methods to create Web applications.
 What Is BizTalk? 357

What Is BizTalk?
BizTalk is a B2B solution that allows you to build integrated applications. The key to
BizTalk is integration. Microsoft designed this product to fill a hole in the application sup-
port currently provided for B2B communication. Their documentation also states you can
use BizTalk for in-house development (a true statement), but this misses the real purpose of
BizTalk: sharing data with other companies. Although you can create simple applications
using BizTalk, the real power of this product doesn’t appear until you delve into the com-
plexities of B2B development.
The Microsoft documentation points out that you can share data using BizTalk regardless of
the programming languages and operating system platforms used by the two parties. The
truth is that BizTalk doesn’t possess any magic that suddenly removes the barriers that devel-
opers have fought for many years. However, BizTalk does rely on standardized forms of com-
munication, including XML and SOAP. The communication that BizTalk provides makes it
special. Many of the files use XML for storage, making them portable to other platforms.

BizTalk supports more than just XML and SOAP; it also supports other standards such
as Electronic Data Interchange (EDI). This Appendix will focus on the protocols and
standards that are relevant to this book. Check Microsoft’s BizTalk site for additional
information on BizTalk features.

Of course, we’ve been building applications throughout the entire book without the aid of
BizTalk. All of these applications communicate with other platforms just fine and for a lot
less money. A SOAP developer would need a compelling reason to use this product. BizTalk
has many other tricks up its sleeve that make it a worthwhile purchase. For example, BizTalk
includes design tools based on Visio that help you create better solutions from the ground App
up. Many developers will choose to use BizTalk for the tools alone.
B
Sometimes the sharing that companies need to do is service related, rather than data related.
For example, an airline could provide access to their reservation system as a service. One
way to advertise such a service is through Universal Description, Discovery, and Integration
(UDDI). BizTalk also provides the means to share services between businesses. While
UDDI is a means to share services publicly with potential business partners that might not
know much about you, BizTalk provides sharing at a private level. The integration is
between companies that know each other well and share a trust relationship.
BizTalk provides the means to describe the interfaces contained within components.
Describing what you want to do to someone else is an essential communication in the B2B
environment. It’s important to consider how best to describe the required interfaces, meth-
ods, and parameters. BizTalk provides tools specifically for this purpose.
358 Appendix B Microsoft BizTalk and SOAP

The type of integration that BizTalk provides implies some level of data translation. When
two businesses share data, they also share data formats. Unfortunately, businesses seldom
consult each other about data formats, so the organization of your data will differ from
those of your partners. Another purpose for BizTalk is to provide seamless data translation.
When a user from Company A sends a form to another user in Company B, BizTalk auto-
matically translates the format of the data so each user sees the data in a familiar format.
This mapping feature is another productivity aid for developers. Instead of writing transla-
tion code by hand, you can use a graphical utility to perform the task. BizTalk Mapper relies
on another standard, Extensible Stylesheet Language Transformations (XSLT), to perform
its task.

Microsoft is currently working on additions to BizTalk for the healthcare industry. These
additions will implement the requirements of the Health Information Privacy Act of
1999 (HIPA). Rumors indicate that Microsoft is working on other additions as well.
You’ll want to check for special additions to BizTalk if you work in an industry with spe-
cial requirements.

As you can see, BizTalk is a product for designing and implementing large-scale
applications. It includes a variety of tools that allow you to design an application with little
effort. The centralized database eases team development and ensures that none of the data
gets lost. BizTalk is one of those tools of the future that you keep hearing about and that
Microsoft hopes you’ll choose as your next development tool.

How BizTalk and SOAP Work Together

BizTalk uses SOAP and other protocols to transfer documents between companies. You can
also choose from Hypertext Transfer Protocol (HTTP), Hypertext Transfer Protocol Secure
(HTTPS), and Simple Mail Transfer Protocol (SMTP) as transport protocols. The point is
that SOAP is one of many available protocols at your disposal when you use BizTalk. This
product gives you a choice of which protocol to use in a given situation.
BizTalk is an extensible architecture. You can create add-ons that enhance its performance
and allow BizTalk to do more than it can out of the box. The extensible architecture means
that vendors can add updates to the current SOAP support as the specification changes (an
almost daily event). Theoretically, this means that you could also see special SOAP tools in
the future (perhaps as part of a planned BizTalk Server Resource Kit).
Like the Microsoft SOAP Toolkit, BizTalk provides full SOAP 1.1 support, along with Web
Services Description Language (WSDL) support. BizTalk can also use older Service
Description Language (SDL) and SOAP Contract Language (SCL) files. WSDL has
replaced these standards in most situations. You’ll only use the older standards with existing
applications or products that vendors haven’t upgraded for the new standard. It’s important
that BizTalk provides this support because newer versions of the Microsoft SOAP Toolkit
and other vendor products don’t.
 An Overview of Useful BizTalk Utilities for SOAP 359

Microsoft provides documentation for SCL at http://msdn.microsoft.com/net/

scl.zip. Even if you don’t use this method of describing your service, the document
makes for interesting reading; you can see where SOAP started and where it is today.

BizTalk also provides a lot of XML support. Because SOAP is essentially a wrapper around an
XML description of a data communication, you can use BizTalk to create new types of SOAP
messages. Of course, the extent to which you can use this capability depends on your business
requirements and the amount of time you want to spend adapting existing BizTalk file formats.

An Overview of Useful BizTalk Utilities for SOAP

This section of the Appendix shows you one of the biggest reasons to download an evalua-
tion copy of BizTalk, at least from the perspective of the SOAP developer. We’ll explore a
few of the tools that BizTalk provides. The following sections won’t tell you everything
about the tools, but they will tell you about some of the interesting things you can do with
them in an effort to create your next SOAP application.

Microsoft plans to introduce a BizTalk Server Resource Kit. However, their exact plans
are unknown as of this writing. You’ll want to watch the BizTalk Web site (listed earlier
in the chapter) for updates and new information about the product. In the mean time,
you can check out downloads Microsoft makes available for this product.

BizTalk Editor
You’ll normally use BizTalk Editor and BizTalk Mapper (described in the next section) App
together. BizTalk Editor allows you to define data formats. The Editor describes formats in
an XML document that you can then import into BizTalk Mapper. You can use the BizTalk B
Mapper to describe transitions between one data format and another. For example, you might
need to translate data created by a partner into a format that your system will understand.
When you start BizTalk Editor, you will see a window similar to the one shown in Figure
B.1. (I’ve already loaded the example PaymentSpec project that comes with BizTalk Server
in this case—your actual display will vary by project type and content.) Note that the left
pane shows an overview of the data for this application. This could easily show the output
from a component in your application. Figure B.1 shows a few of the levels expanded so that
you can see the relationship between entities. Each of the tabs in the right pane allows you
to define the attributes of the selected element. For example, the Name element is an
attribute type and is of the string data type.
You can create a number of project types with BizTalk Editor, including blank specification,
XML, Electronic Data Interchange for Administration Commerce and Transport (EDI-
FACT), and X12. X12 is the American National Standards Institute Accredited Standards
Committee electronic data interchange standard. SOAP developers are most likely to use
360 Appendix B Microsoft BizTalk and SOAP

the XML project. Figure B.2 shows the XML document specifications you can create.
Notice that Figure B.2 shows a Simple SOAP Envelope project.

Figure B.1
The BizTalk Editor
helps you define
schemas.

Figure B.2
You can create a
number of XML
document types with
BizTalk Editor.

When using the Simple SOAP Envelope project, you’ll begin with a project that includes an
envelope as a top element, a header, and a body. All you need to do is create a SOAP mes-
sage as if you’re the application. Figure B.3 shows an example of the document you might
create for the simple component example in Chapter 4, “Using SOAP to Create a Simple
Application.” Note that using this technique describes the desired results first, rather than
concentrating on the means to obtain the desired result.
Using an XML project could save you time in other areas of application development. For
example, you could use the resulting file for comparison purposes with automatically gener-
 An Overview of Useful BizTalk Utilities for SOAP 361

ated WSDL and as input to a low-level API class. In short, BizTalk Editor can provide
description services for any part of your application, not just as part of the application design.

Figure B.3
An example of a SOAP
envelope created with
BizTalk Editor.

One of the most interesting features of BizTalk Editor is creating a specification and out-
putting it in XDR format. Remember that this is one of the inputs required for SMO in
Visual Basic. All you need to do is use the Tools | Export XDR Schema command to display
the dialog box shown in Figure B.4. Make sure you give the output file an XDR extension or
SMO won’t recognize it. Note that BizTalk Editor defaults to XML as the output extension.
App
Figure B.4
Use the BizTalk Editor B
to create the XDR files
you need for SMO.
362 Appendix B Microsoft BizTalk and SOAP

Figure B.5 shows an example of the XDR file output from BizTalk Editor. Interestingly
enough, although this file does work with SMO, it displays an error message in XML Spy. This
points to potential XDR formatting problems by BizTalk Server. Consequently, there’s no guar-
antee that the XDR files you create with BizTalk Editor will work with other vendor products.

Figure B.5
An example of the
XDR output from the
BizTalk Editor.

As with most XML editor tools, BizTalk Editor allows you to perform validation on your
design. Simply use the Tools | Validate Specification or the Tools | Validate Instance com-
mands to get the job done. You’d use the Validate Specification version for validation of the
currently displayed file. Note that any warnings will appear on the Warnings tab on the bot-
tom of the display show in Figure B.1—you won’t receive another form of notification.

You’ll find the BizTalk projects, XDR output, and SMO project used for demonstration
purposes in this Appendix in the \Appendix B directory of the source code available
from the Que Web site. The source code is at http://www.quepublishing.com.
Depending on how you set up your system, you should be able to compile the
SMO-generated component and use it to access the AddIt component discussed in
Chapter 4. Make sure you modify any server-specific information as required for
access.

BizTalk Editor supports records and fields as major data divisions. These two object types
have different representations in the left pane of the BizTalk Editor so that you can see them
with ease. Generally, I find that using records for major document constructions—such as
the envelope and body—and using fields for data input—such as Add1 in the example—
allows you the greatest flexibility. Notice that all of the Microsoft-supplied projects also use
this node and leaf setup.
 An Overview of Useful BizTalk Utilities for SOAP 363

BizTalk Mapper
After working on several SOAP applications and discovering data formats are such a prob-
lem, the reason to use BizTalk Mapper became clear. This utility allows you to match a
source specification to a destination specification using a graphical tool rather than pencil
and paper. Figure B.6 shows an example of mapping the AddIt component from Chapter 4
to a fictitious partner’s format. The two methods exchange about the same information, but
use different method calls and data formats to do it.

Figure B.6
The BizTalk Mapper
makes it easy to con-
vert data from one
format to another.

App
Note the dark lines in Figure B.6. Those lines show the connection between the two specifi-
cations. Because Add1 is an integer data type and Input1 is a string, the example needs to
B
provide some means of conversion between the two. Note that there’s a block in the middle
area, known as a functoid. You use a functoid to create conversion routines between two docu-
ment specifications. Figure B.7 shows the Functoid Palette that you can access by right-
clicking the space between specifications and selecting Functoid Palette from the context
menu.
Although BizTalk does provide a number of predefined functoids, they won’t cover every
need. In this case, the example requires a custom functoid to perform the required conver-
sion. The Advanced tab of the Functoid Palette dialog box contains a Scripting Functoid
that you can drag and drop into the gray area between document specifications.
364 Appendix B Microsoft BizTalk and SOAP

Figure B.7
The Functoid Palette
contains a number of
predefined routines,
including several for
converting numbers.

Right-clicking a Scripting functoid and choosing Properties will display a Functoid

Properties dialog box. The General tab of this dialog box will allow you to add, delete,
promote, and demote parameters. The Script tab shown in Figure B.8 allows you to create
conversion routines using VBScript. The BizTalk documentation isn’t clear about which
scripting commands you can use, but none of the VBScript functions I tested failed.

Figure B.8
Functoids are actually
scripts used to
convert data.

Saving a BizTalk Mapper document automatically compiles it and displays the Warnings tab.
In most cases, you’ll see a message saying the XSL compiler experienced 0 errors. In addi-
tion, you can view the actual output of the BizTalk Mapper on the Output tab shown in
Figure B.9. Notice that BizTalk Mapper saves any functoids you create as CDATA sections,
as shown in the figure. It pays to spend time looking at the output to ensure you understand
how BizTalk Mapper implements certain operations.
Interestingly enough, you can use the Tools | Test Map command to check the validity of
your map. You’ll see the results of test data on the Output tab. The information isn’t quite
complete, but it shows a good simulation of your map, as shown in Figure B.10.
 An Overview of Useful BizTalk Utilities for SOAP 365

Figure B.9
BizTalk Mapper saves
functoids as CDATA
sections in the result-
ing XML file.

Figure B.10
Test your map to
ensure you receive the
intended output.

App

Because BizTalk Mapper compiles the results of your graphing operations to an Extensible B
Stylesheet Language Transformations (XSLT) file, you can use it outside of BizTalk Mapper.
The process of saving your map compiles the design automatically, but it doesn’t save the
external file. To do that, you need to use the File | Save Compiled Map As command.

BizTalk Orchestration Designer

One of the centerpieces of the BizTalk suite is BizTalk Orchestration Designer. Anyone who
has used Visio before will be familiar with this utility as well. All that BizTalk does is add a
BizTalk interface over the existing Visio product. Figure 8.11 shows a typical BizTalk
Orchestration Designer window with one of the sample BizTalk application business
processes loaded.
366 Appendix B Microsoft BizTalk and SOAP

Figure B.11
The BizTalk
Orchestration
Designer relies
on Visio to perform
its work.

As you can see, the standard display includes two tool panes and one drawing pane (center).
The left pane contains flowchart tools. All you do is draw the steps that the user or other
entity will follow fulfilling the task in question. The flowchart maps the business process
used to create or exchange data. Of course, this is the SOAP view of the BizTalk
Orchestration Designer functionality—real world applications will use this product for
much more.
The right pane contains implementation tools. This pane addresses the question of how you
anticipate fulfilling the tasks spelled out by the process flow. Obviously, the symbols shown are
for a wide range of business processes. SOAP developers can still use the component symbol
to show the interactions that take place within the SOAP portion of a larger application.
BizTalk Orchestration Designer divides the grid, or drawing pane, into three areas. The left
side contains the business process that you’re describing. The right side contains all of the
implementation details. A “no man’s land” in the center delineates the business process and
implementation. The center section also shows connectivity between the component and
the process that uses it.

Unlike many other BizTalk utilities, the BizTalk Orchestration Designer uses a special file
format to store information, instead of a standard format, such as XML. The XLANG file
format looks similar to XML, but it really isn’t. Most developers will never need to learn
XLANG, but it pays to know that BizTalk Orchestration Designer uses this special file
format in case you need to edit the file manually.
 An Overview of Useful BizTalk Utilities for SOAP 367

Figure B.12 shows a simple implementation of the AddIt example from Chapter 4. This
example shows how many of the BizTalk Orchestration Designer symbols work.

Figure B.12
A simple diagram
showing the imple-
mentation of the
AddIt example from
Chapter 4.

Every diagram begins with the supplied Begin symbol. Joins and forks allow you to create
multiple paths of execution or close logic paths, as shown in the figure. You’ll use actions
and decision symbols as you would in any application.
Notice the port connection in the center of the diagram. This is how BizTalk expresses con-
nectivity between a process and its implementation. In this case, we’re connecting to the
AddIt component using the AddNumbers interface.
The various wizards make it possible to view how the application will interact with the com- App
ponent during the description process. For example, you’ll use the Class Information dialog
box shown in Figure B.13 to locate the component and interface that you want to use within
B
the application. (If you want to use multiple interfaces, you’ll need to define more than one
port.) The Class Information dialog box is one of three dialog boxes that appear as part of a
wizard that defines a COM component’s method of binding when you add a COM
Component symbol to the drawing. The other two dialog boxes define the type of connec-
tion (static, dynamic, or not instantiated), security, and transaction support.
You have to define a method of communication as part of creating a connection from an
action to a port. BizTalk Server supports both synchronous and asynchronous communica-
tion. The Message Specification Information dialog box shown in Figure B.14 allows you to
choose from one of the methods the component supports. Notice that it divides input from
output fields. This is your chance to verify the component will react as anticipated.
368 Appendix B Microsoft BizTalk and SOAP

Figure B.13
Adding a COM compo-
nent symbol starts the
wizard that displays
this dialog box.

Figure B.14
BizTalk Orchestration
Designer verifies input
and output for the
component method
you choose.

The BizTalk Orchestration Designer is far more complex than this short section indicates.
However, this section does provide you with the SOAP developer’s overview of the product.
You’ll want to explore BizTalk in more detail if you plan to spend much time working with
complex B2B applications.

BizTalk Fixes a Few SOAP Problems

Throughout the book, we examined various problems with SOAP. For example, you know
that toolkits from different vendors don’t necessarily interoperate well. Although the SOAP
specification provides a description of features such as attachments, none of the existing
toolkits provides this support. In short, SOAP toolkits have a long way to go. However,
BizTalk does provide a few solutions to your SOAP woes.
 Is BizTalk the SOAP Add-On for Your Company? 369

One way that BizTalk improves on the capabilities of native SOAP is by providing guaran-
teed one-time delivery. This isn’t the same as providing transactions, but it does come close.
A guaranteed one-time delivery means that you have fewer worries that the document will
disappear in transit. However, this guarantee doesn’t include damage to the file. In other
words, the document will arrive at the other end, but there isn’t any guarantee that the data
it contains will remain safe.

You can learn about BizTalk in a variety of places. Two of the more interesting Web sites
are Microsoft Business (http://www.microsoft.com/business/default.asp)
and BizTalk.org (http://biztalk.org/home/default.asp). If you need help using
BizTalk, check out the wealth of newsgroups under microsoft.public.biztalk.
(These newsgroups may only appear on the news.microsoft.com news server.)

SOAP currently has security problems. You can secure a message using an add-on such as
secure sockets layer (SSL), but some developers feel that this isn’t enough—they want the
security built into the specification. BizTalk can’t really fix this problem because it is a mat-
ter for standards organizations to decide. It does make security easier to use. You’ll find that
BizTalk includes both SSL and Secure/Multipurpose Internet Mail Extensions (S/MIME)
support. BizTalk supports S/MIME version 1.0, 2.0, and 3.0 as part of the package.
Another interesting addition to BizTalk is MIME support. You could use this support to
build SOAP applications with attachments today. However, you’d want to have more reason
that the benefit of attachments to buy this product.
BizTalk relies heavily on Queued Components. This means you can create disconnected
applications that communicate using a queue, rather than a direct connection. Using
BizTalk means that you can send purchase orders to a queue and allow the server to handle
App
them during periods of low activity.
B
Is BizTalk the SOAP Add-On for Your Company?
At this point, you need to ask whether BizTalk is the B2B solution that you need today. For
many businesses, the answer is no. BizTalk may provide an interesting platform for experi-
mentation, but some businesses aren’t ready for the robust environment it provides. BizTalk
represents a major change in the way companies handle data and transfer it to their part-
ners.
If you are unsure whether BizTalk is the right solution, you can look at the issue in another
way. The real question is whether you need to transfer data or whether you actually need to
translate and manage it. If all you need to do is transfer data, then SOAP is probably the
best technology to use.
370 Appendix B Microsoft BizTalk and SOAP

BizTalk is the SOAP programmer’s dream in many ways. Developers who spend all of their
time working on Web applications will appreciate the tools that BizTalk provides. This
product makes easier to design and implement complex systems quickly.
You’ll also enjoy the SOAP features that most SOAP toolkits don’t provide today, such as
the ability to use attachments. Some developers might get BizTalk for the tools and capabili-
ties it provides for translating documents alone. In fact, when you perform the product
installation, you’ll find an option to create such a setup. Microsoft intended BizTalk to fill
the gaps in the developer toolkit.
The feature that many developers like best is data translation. In Chapter 8, “Providing
Remote Database Access,” you saw what the typical SOAP programmer has to do to make
even simple data transfers work well. Imagine the amount of work that developers perform
to implement data transfers for the large systems that companies have in place right now.
Any form of automation is welcome, and BizTalk does a terrific job in this area.
This Appendix only demonstrated the smallest portion of the capability that BizTalk pro-
vides. It’s an industrial-strength solution to the problem of B2B communication. So, is this
product overkill for your company or just the right solution at the right time?
 APPENDIX
C
Third-Party Tool Reference
In this appendix
Finding the Right Tools 372
Master 2.0 374
MZTools Add-In for Visual Basic 376
psWSDL Wizard 381
tcpTrace 383
XML Spy 384
372 Appendix C Third-Party Tool Reference

Anyone who creates applications develops a set of specialty tools that allows them to work
faster and with fewer errors. The third-party tool market is strong because the development
package that arrives on your doorstep seldom contains a complete set of tools for every
anticipated need. I ran into the problem while writing this book and working on a small
consulting project. In some situations, I found that getting a third-party tool helped me
develop code faster. In other cases, a third-party tool proved indispensable for locating and
squashing bugs in my code (yes, I generate lots of the critters, just as anyone else does).
The first part of this appendix contains some of the tricks I use for finding new tools.
Sometimes I’ll find an area of a program that I can’t solve by reading an article or book,
and getting online doesn’t help because no one else is running into that particular problem.
That’s when it’s time to sit back and look at the problem from a distance—take the big view
of the application as a whole. In many cases, I’ll determine that I need a new tool to conduct
further research into the problem. That’s when I start hunting for new tools online if my
toolbox lacks a tool that does what I need it to do. Someone has likely run into the same
problem you ran into before and developed a tool to deal with it. If you can’t find something,
perhaps it’s time to develop a tool of your own and make it available for everyone else to use.
This rest of this appendix contains the list of development tools that I found most helpful.
This appendix isn’t intended to provide a full list of every third-party tool on the market,
partly because I probably haven’t seen all of the available third-party tools. You should con-
sider this list the tools that I find essential for my toolkit and some potential additions to
your toolkit. If you have a tool that you feel is indispensable, write me at [email protected]
and let me know why you think this tool is so great. Once I have a chance to use the tool
myself, I’ll include a write-up about it on my Web site at http://www.mwt.net/~jmueller/.

Finding the Right Tools

Before I begin the summary of tools in this appendix, I wanted to spend a few minutes talking
about finding the right tool for the job. You categorize tools by type. I use two basic types of
tools: those that allow me to create applications faster and those that help me fix bugs.
I use different search techniques for each tool type. The following sections tell you how I
found the tools in this appendix and what you can do to make your searches more produc-
tive. After all, having a full toolbox often means the difference between getting a project
done on time and fiddling with it long after the delivery date.

Tools That Make You Productive

Magazines and the trade press are the two best places to find productivity tools. If I’m looking
for a tool, I’ll skip the news and go directly to the columns and reviews. Columns are the
best source of free utilities. The author will commonly mention the free utility as part of the
current topic. Often you can simply scan for the special font a magazine uses for URLs or
look for the familiar “http” portion of the URL.
 Finding the Right Tools 373

Some magazines also provide large online search engines for finding utilities. For example,
I often find new productivity tools on the Ziff-Davis Web site (http://www.zdnet.com/
downloads/), Open Source Development Network (http://www.andover.net), or C/NET
(http://download.cnet.com/downloads/). The problem with these resources is that you’ll
spend a lot of time wading through downloads that don’t suit your needs before finding the
one that will.
Another good place to look for productivity tools is vendor Web sites. Vendors want you to
become productive with their tools because that generates good publicity. Microsoft is one
of the best examples of a vendor who provides good third-party product coverage. Microsoft
will often provide entire pages of third-party product listings. A vendor employee normally
works with these products and only lists those that perform well.
The places of last resort for finding productivity utilities include standard search engines
and newsgroups. I’ll use a search engine such as Dogpile (http://www.dogpile.com/) or
Google (http://www.google.com/advanced_search) when looking for utilities. They tend
to provide more hits than a search engine like Hotbot (which I use for many other types
of research).
Using a newsgroup to research productivity tools is a little like playing Russian roulette—
you never know when the next tool will actually damage your project, rather than help you
write it faster. I find that vendors inundate me with advertisements of dubious quality when
I ask a question about utilities on a newsgroup. In addition, many of the newsgroup users
are looking for the same utilities that you are. They need the same answers and don’t know
where to find them.

Tools That Fix Problems

Sometimes you can’t solve a problem using the standard fixes because you lack the informa-
tion needed to understand the problem fully. That’s where a diagnostic utility comes into
play. A diagnostic utility allows you to look at your application and its data in a new way or to
ferret out some new detail about your system. Of the two types of utility, developers always
need more diagnostic utilities. In fact, this is normally the first tool that a developer acquires.
Unlike productivity utilities, newsgroups are great places to find out about diagnostic
utilities. People normally spend time on newsgroups to help others or to discuss problems App
they’re having with their own application. In many cases, you’ll hear how well a utility
works in actual use without the embellishments that sometimes appear in magazines. C
Another place to look for diagnostic program listings is in the vendor’s frequently asked
question (FAQ) listing. This list can appear in a number of places. For example, Microsoft
has a place on its Web site for its Knowledge Base. You can also get a subscription to
Microsoft Developer Network (MSDN) and gain access to a searchable FAQ list. Many
of the articles point you toward special tools that Microsoft doesn’t include in their other
product packages. Of course, many of these tools are unsupported, so getting help with
them is nearly impossible unless other developers have used the tool.
374 Appendix C Third-Party Tool Reference

Third parties often develop diagnostic utilities, but not the same ones who create productiv-
ity utilities. Developers often create the diagnostic utilities out of sheer frustration. They’re
tired of working without the proper tools. These grass-roots effort tools have a few rough
edges, but normally help locate problems and that’s all that counts.
Sometimes you’ll have to create your own tools. I won’t spend a lot of time discussing that
route here, but it’s important to realize that you sometimes need to create a tool before you
can create an application. If you ever develop a tool that you feel is especially helpful, let me
know and I’ll give you a plug on my Web site after checking it out. SOAP development is so
new that there’s plenty of room for innovation.

Masker 2.0
Security is a major concern for SOAP developers. Some developers are using Secure Sockets
Layer (SSL); others are waiting for the World Wide Web Consortium (W3C) to complete
work on the XML encryption specification (http://www.w3.org/Encryption/2001/). A few
developers are trying other techniques to hide their data in an attachment, then send the
attachment using the SOAP attachment technique (see the “Understanding SOAP
Attachments” section of Chapter 2 for details).
Masker (http://www.masker.de/) is a product that helps fulfill this last data encryption need.
It permits you to create a package containing your data using a carrier file technique. The
carrier file is a BMP or other graphic file (GIF, JPG, and TIF). You can also use certain
program (EXE and DLL), video (AVI, MOV, and MPG), and audio (WAV, MID, SND, and
MP3) file formats. The carrier file retains its current functionality, so you could send a GIF
file containing your company logo. As far as anyone is concerned, the logo is simply part of
a message delivered using some method. Only those who have the proper password can
open the encrypted data contained within the carrier file.

Masker is one of many carrier file data protection products on the market. However,
not every product includes encryption—they rely on the hidden nature of the carrier
file for data protection. That’s why Masker is special. It protects your data using RC4
encryption, along with the hidden nature of the data within the file.

Using Masker is easy. I use the following steps to secure data that I want to send using a
MIME attachment with SOAP. (The procedure assumes you have already installed Masker
and have it running on your system.)
1. Use the File | New Carrier File command to display the dialog shown in Figure C.1.
2. Type the name of the carrier file. You can also use the browse button to locate the
carrier file on the hard drive.
3. Select one or more files to hide within the carrier file using the Add button. You can
also highlight a file and remove it from the carrier using the Delete button.
 Masker 2.0 375

Figure C.1
Masker allows you to
use any of the sup-
ported file formats as
a carrier file.

4. Type a description for the carrier file.

5. Type a password, then confirm the password. I used a password of “hidden” for the
example file found on the CD, but you should use something a little more secure. I
usually use two common words at least four characters long separated by a special
character or number for medium security. You can use randomly generated strings
of letters, numbers, and special characters for a higher level of security.
6. Choose a write protection option. Clear the option if you want to allow others to
modify the carrier file after Masker writes it. The default option write protects the file
so that no one else can modify it accidentally.

Masker and other carrier file products rely on Windows streams to hide data. A file
can contain multiple streams, but only the first stream is readily accessible using most
applications. Since these applications are unaware of the hidden streams within the App
file, using them to modify the file will destroy the hidden streams containing your data.
In short, allowing someone to modify the original copy of a carrier file with a standard
C
application is an almost certain method for losing the data it hides.

7. Choose a delete option. Masker clears this option by default so that it won’t erase your
original file from the hard drive after adding it to the carrier file. Given the objective of
using this product for SOAP is to provide a secure means of moving the file from one
point to another, I would recommend leaving the option cleared. However, if you won’t
376 Appendix C Third-Party Tool Reference

need the file on your local hard drive and are certain the remote party will know how
to handle the Masker file, you can always delete the original on the hard drive for
added security.
8. Click Hide. Masker will hide the file within the selected carrier. You’ll see a summary
window such as the one shown in Figure C.2 when the process is complete.

Figure C.2
Masker shows a
summary of the files
hidden in the carrier.

If you opened the carrier file using a standard utility at this point, all you would see is the
original. Masker completely hides the data contained within the carrier and you can only
view it within Masker after supplying the proper password. The only indication that the file
changes is that the file size changes by the amount of the data that you store within the
carrier, plus a little space for housekeeping information.

MZTools Add-In for Visual Basic

Developer productivity begins with the IDE. Many developers complain that Visual Basic lacks
some of the IDE menu essentials they’d like to see. MZTools (http://www.mztools.com/) pro-
vides a freeware addition for Visual Basic that could truly revolutionize your way of working
with this product. Finally, you’ll have the menu entries that you originally wanted to see in
Visual Basic. This product also enhances your ability to modify the Visual Basic IDE to suit
your particular tastes.
Adding MZTools to the Visual Basic IDE is easy. Just download the file from the MZTools
Web site, extract the files to the directory you want to use to store the product, and use
RegSvr32 to register the product. (Type RegSvr32 MZTools2.DLL at the command line
prompt to register the DLL.) The tool should automatically load and startup the next time
you open Visual Basic (you can always manually load it using the Add-In Manager as shown
in Figure C.3).
MZTools will also add a new toolbar to the Visual Basic display that provides access to all of
the features. Figure C.4 shows what the toolbar looks like. The following list provides a
brief description of each toolbar button from left to right.
 MZTools Add-In for Visual Basic 377

Figure C.3
MZTools should load
automatically, but you
can also load this tool
manually using Add-
In Manager.

Figure C.4
MZTools provides
easy access to its
features through
a new Visual Basic
toolbar.

■ TabIndex Assistant: Allows you to change the tab order of controls within an
application. Figure C.5 shows a typical example from Chapter 5, “Migrating an
Application from DCOM to SOAP,” of the book. All you need to do to change the
tab order is select the control in question and use the arrows to move it up or down
in the list. MZTools assigns controls at the top of the list lower numbers beginning
with 1.

Figure C.5
The TabIndex
Assistant allows you
to change the tab
order of controls by
moving them around App
in the list.
C
378 Appendix C Third-Party Tool Reference

■ Review TabIndex Property: Verifies that the controls on a form are in a logic order
using the same algorithm that TabIndex Assistant uses for automatic ordering. This
feature works fine in most cases, but can provide dubious results on complex forms.
Figure C.6 shows the output dialog for this tool.

Figure C.6
Review TabIndex
Property performs
analysis of your form
and offers suggestions
for reordering the
tab stops.

■ Find: Displays a Find dialog that helps you to search for code at the project group,
project, or module level. The advantage of this option is that you can perform the
search at any time instead of only when editing code. Once you have found a string
within the code, you can replace all occurrences of that string in all modules at the
same time. This two-step approach is much faster than Microsoft’s method of replacing
strings one module at a time. Of course, you need to use it carefully to avoid shooting
yourself in the foot.
■ Add Procedure: This button sounds simple, but it combines a lot of functionality into
one dialog (shown in Figure C.7). Consider this a more convenient superset of what
Visual Basic has to offer. You can use it to add more than just procedures and functions.

Figure C.7
The Add Procedure
dialog is a convenient
superset of the stan-
dard Visual Basic
offering.
 MZTools Add-In for Visual Basic 379

■ Add Module Header: Adds a simple heading to the top of the module that includes
the name of the associated form, the filename, the name of the author, and the date. All
you need to fill in is the purpose of the module. You can modify the module header
contents on the Module Header tab of the Options dialog.
■ Add Procedure Header: Adds a simple heading to the beginning of a procedure that
includes the procedure name, the author name, and date. You need to include the pur-
pose, result, input parameters, and output parameters for the procedure. (The program
does create the blank entries for you.) You can modify the procedure header contents
on the Procedure Header tab of the Options dialog.
■ Add Error Handler: Adds an error handler header, plus an example error handler.
You’ll need to add the error handler code you want to modify the default entries. You
can modify the error handler code and associated comment on the Error Handler tab
of the Options dialog.
■ Add Both: Adds both a procedure header and an error handler. Combine this with
the Add Procedure button to create procedure stubs at the beginning of a program-
ming session.
■ Procedure Callers: Highlight a procedure name within the code, then click this button
to see a display similar to the one shown in Figure C.8. Procedure Callers will help you
find every caller for a particular procedure.

Figure C.8
Procedure Callers
helps you locate
every procedure
that calls on the
procedure you high-
lighted before
clicking the button.

App
■ Review Collections: Checks the collections within an application to ensure they adhere
to certain standards (see vendor Web site for details). MZTools considers every class C
that implements the NewEnum or Item methods a collection. The results of the review
appear in a Window similar to the one shown in Figure C.8.
■ Clear Immediate Window: Removes text from the Immediate Windows, which is a
handy feature during debugging sessions.
■ Statistics: Displays the current statistics for your application as shown in Figure C.9.
You can sort the statistics to learn about your code. For example, you may want to find
out who includes the most number of comments in their code when working in a team
environment. MZTools excludes blank lines from the statistics.
380 Appendix C Third-Party Tool Reference

Figure C.9
The Statistics
dialog can tell
you a lot about
your coding habits.

■ Task List: Displays a list of the current tasks within your application as shown in
Figure C.10. Notice that you can give your task a priority, assign it to a specific person,
and provide both a long and short description. The only feature missing is to assign the
task to a specific line or area of code.

Figure C.10
The Task List allows
you to track and
prioritize current
project requirements.

■ Copy Controls with Code: Copies not only the control (with all settings intact), but
the associated code as well. This means that you can move buttons that you commonly
use from project to project with relative ease.
■ Paste Controls: with Code Pastes a control that you copied using Copy Controls with
Code onto the current form.

The MZTools Web site includes a Delete Controls with Code button, but the version of
the product that I used didn’t include this feature. It may appear with the next version
of the product.

■ Add Line Numbers: Adds line numbers to your listing, making it easier to find errors.
■ Remove Line Numbers: Removes the line numbers added using the Add Line
Numbers feature.
■ Reload from Disk: Reloads the current project from disk without saving any changes.
MZTools does display a warning message before it performs this task.
■ Close Windows: A useful aid for cleaning up your work area, this feature closes all
open windows.
■ Options: Displays a dialog that allows you to change the MZTools options. Figure
C.11 shows what this dialog looks like. As you can see, it allows you to change all
application features. You can even use this dialog to assign shortcut keys to the various
MZTool features.
 psWSDL Wizard 381

Figure C.11
The Task List helps
you track and priori-
tize current project
requirements.

psWSDL Wizard
Most of the examples in the book required the generation of a WSDL file. Since this file is
unique for every Web server, you’ll likely find that you need to generate many WSDL files as
you create SOAP applications. Some people may find that they don’t like the Microsoft ver-
sion of the tool. The psWSDL Wizard from Phalanx Systems (http://www.phalanxsys.com/
soap/wsdlwiz.htm) represents an alternative to the Microsoft option.

Using this wizard is relatively easy. You follow the five-step process listed below (I’ll assume
you have the psSOAP WSDL Wizard installed).
1. Start the application. Click OK to get past the first dialog. You’ll see a dialog similar to
the one shown in Figure C.12.

Figure C.12
psWSDL Wizard looks
much like the
Microsoft offering
when you start it up.

App

2. Type the name of a DLL to use for the WSDL document, the WSDL file you want to
create or change, and an URL for the SOAP listener. Click Next. You’ll see a hierarchical
display of the component’s file, class, and interface as shown in Figure C.13.
382 Appendix C Third-Party Tool Reference

Figure C.13
Choose the classes
and interfaces you
want to use from
this display.

3. Select the classes and interfaces you want to use for your application. Click Next. You’ll
see the WSDL Service Setup dialog shown in Figure C.14. (Not all options appear high-
lighted in the figure.) This is where the psWSDL Wizard differs from the Microsoft
offering. Notice that you can select a specific method and choose the manner in which
the application will interact with the component. If you right click a method, you’ll see
all the by reference arguments. You can choose whether these arguments are input only
or in/out. This flexibility allows you to create better WSDL files, which in turn permits
you to produce better application in less time.
4. Set the WSDL options. Click Next. You’ll see a dialog that asks for the WSDL defini-
tion and service name. Supply values that you want to use within your application to
access the WSDL file.
5. Type a WSDL definition value and WSDL service name. Click Finish. PsWSDL
wizard will write the file for you.

Figure C.14
You can choose how
an application will
interact with your
component from
this dialog.
 tcpTrace 383

The output from psWSDL Wizard is different from the standard Microsoft SOAP toolkit
utility. You won’t see a WSDL/WSML file pair in this case. In addition, this utility won’t
generate an ASP or ISAPI listener for you. However, you do receive a full-fledged WSDL
file with all type information in place. In some respects, the use of this single file is superior
to the Microsoft combination. This file will work with the Microsoft SOAP Toolkit. I didn’t
test it against other SOAP toolkits.

tcpTrace
Many of the SOAP toolkits on the market don’t include any form of trace utility. You need
this utility to track the messages to and from the server. Analyzing these messages helps you
detect certain types of bugs in your application. Even though the Microsoft SOAP Toolkit
does include a trace utility, it doesn’t provide any means for creating a log file on disk—a
prerequisite if you want to track the messages generated by a large application. pocketSOAP’s
tcpTrace (http://www.pocketsoap.com/tcptrace/) addresses both of these problems.
Using tcpTrace is relatively easy. You’ll see the dialog shown in Figure C.15 when you start
the application. The three blanks allow you to configure source and destination communica-
tion parameters. I’ll describe how these parameters work in the paragraphs that follow.

Figure C.15
A tcpTrace session
starts by configuring
the source and desti-
nation parameters.

tcpTrace assumes that you’ll use localhost as the starting point for your communication. What
this means is that instead of pointing your application toward the Web server as normal, you’ll
point it toward localhost (that is, http://localhost/myapp.wsdl). When your application sends
a message, tcpTrace receives it instead of the Web server. tcpTrace sends the information to the
actual site after recording the client information. Likewise, the Web server sends the response App
to tcpTrace, which sends the information back to your application after it records the server
response. The first blank allows you to configure the port that tcpTrace listens on. The second C
and third blanks allow you to set the destination URL and port.

You may have to regenerate the WSDL file if you want to see both sets of messages
between client and server. The WSDL file needs include a pointer to localhost, rather
than to the actual server, so that tcpTrace can intercept the message. You’d need to
perform this same task when using the Microsoft utility. In both cases, the client needs
to know to send its data to tcpTrace and not directly to the server.
384 Appendix C Third-Party Tool Reference

Click OK after you configure these parameters. Now tcpTrace will wait for your client
application to send a message to the server. Figure C.16 shows a simple communication.
Notice that there are two messages. The client makes an initial request and the server
responds with the WSDL file containing the component schema. The client then
provides the information to use the component and the server responds with
component output.

Figure C.16
The tcpTrace output
will show all commu-
nication between
client and server.

XML Spy
XML Spy (http://www.xmlspy.com/) is an excellent editor for XML, XSLT, XDR, HTML,
XHTML, ASP, and a variety of other files. Figure C.17 shows what the main display looks
like with an HTML file loaded. When you look at the output from the product, you’ll
notice that all of the entries are color-coded.
One of the things that impressed me immediately was all of the add-ons you can get for
XML Spy at no cost. The installation program shows you a Web site filled with these
add-ons. If you choose not to install any of the add-ons, XML Spy will prompt you later
when you want to use a feature that you didn’t install. XML Spy is one of the few prod-
ucts that I’ve tried that actually works correctly when it comes to add-on features. I tested
this functionality with scripting support. The entire process worked flawlessly.
 XML Spy 385

Figure C.17
XML Spy works with a
variety of file types,
including HTML.

Typical File Views

The display looks similar to the Visual Studio display. On the left side is a view of your
current project. (Since I’m only using the product to edit files now, there isn’t any project
listing in the figure.) Below the Project window is an Information window used to display
messages. The right side of the display contains three panes called Entry Helpers. The
first pane contains a list of available tag names in this case. The third pane contains a list
of special characters.
XML, XSLT, and XDR files are where this product excels. Figure C.18 shows an example of
an XDR file. Notice that the display shows the various XDR elements arrayed in tabular
format. These tables form a hierarchical display of sorts. You can increase and decrease the
information display by clicking on arrows next to each level of the hierarchy.
Notice that the Entry Helpers contain new entries that match this type of file. The first
pane contains a list of entries, the second a list of attributes, and the third a list of special App
characters. Notice that you can choose to insert or append the new information. You can
also add the information as a new child entry. C
The Information window is also active (lower-left corner). In this case, it contains name of
the selected attribute, along with its type. The window also contains a list of choices for this
type of entry (empty, textOnly, etOnly, or mixed).
386 Appendix C Third-Party Tool Reference

Figure C.18
The use of tables in a
hierarchical display
makes editing XDR
files easier.

The toolbar at the bottom of the display allows you to add new pieces to the XDR file with
little effort. In addition to attribute and element, you can also choose to insert or append
various types of text elements including comments and CDATA.

Viewing Data in More Than One Way

XML Spy provides for a variety of views. Figure C.19 shows a text view of the same XDR
file in Figure C.18. Using this view can help you see the various data relations easier in
some situations. You’ll also find that XML Spy provides a schema View and a browser view.
You can edit the file in any of these views, which helps you compare files in various ways to
determine how to create new documents of the same type. (This is one way that I began
learning about the XDR files used by the Microsoft SOAP Messaging Object Generator.)
You’re not limited to working with local files when using XML Spy either. I was able to
enter URLs for any Web site and view the content from within the editor. Figure C.20
shows part of my Web site. As with every other document that XML Spy reads, you can
view online documents in a variety of ways. I was able to read this Web site in both text
and browser view. A look at other Web sites allowed me to view documents in all of the
four available views. In short, using XML Spy to check out other Web sites isn’t the same
as using a browser; you get a lot more in the way of information.
 XML Spy 387

Figure C.19
The inclusion of
various views makes
using XML Spy easier.

Figure C.20
Even online content is
completely accessible
with XML Spy.

App

C
388 Appendix C Third-Party Tool Reference

Special Features
One of the features I really appreciate is that XML Spy will verify your document for you.
This is important because any error in these text files will cause problems within your appli-
cation. The lack of debugging tools makes some errors particularly hard to find. The only
problem with the error checking features of this product is that it only tells you there’s a
problem—not where the problem is. This could become an issue for developers with larger
files. You don’t want to spend an entire day going through a file line-by-line looking for the
little problem that’s causing the program to squawk. Even so, knowing the error exists is a
big plus.
XML Spy provides the capability of saving, sending, and printing your work in a variety of
ways. I found I used the File/Send by Mail command regularly to send pieces of my project to
other people for review. What the other person will receive is the document in its default for-
mat. Hopefully, they have XML Spy as well, so they can view the document in other formats.
This product actually has too many features to review in a short section of an appendix.
There isn’t room to go into various conversion features, the use of schemas, and generation
of document type definitions (DTDs). You can even create documents in Microsoft Word
and ask XML Spy to import and convert them for you.
 APPENDIX
D
SOAP for Visual C++ Developers
In this appendix
Introduction 390
An Overview of the 4S4C SOAP Toolkit 391
An Overview of the Application 394
Creating the Server-Side Component 395
Creating the Client 400
Handling SOAP Errors 404
390 Appendix D SOAP for Visual C++ Developers

Introduction
This appendix looks at SOAP from the perspective of the Visual C++ developer. It’s impor-
tant to realize that SOAP itself isn’t language specific. You can use SOAP with any language
and operating system that has a toolkit. For that matter, if you don’t have a problem with
writing your own low-level routines, you don’t even need a toolkit. All you need to do is
ouput and accept text in a very specific format. Theoretically, this means you could write
SOAP applications using any language, but using a toolkit is much easier.
At the time of this writing, the Microsoft SOAP Toolkit provides Visual C++ support. The
level of support is on par with Visual Basic, even if the toolkit does contain far more Visual
Basic examples. Microsoft will likely fix this oversight by the time they release the toolkit to
the public. The Visual C++ support has the same limitations that we discussed throughout
the book for Visual Basic. You’ll find the same interoperability problems and the same lack
of support for handheld devices. In short, if you’re a Microsoft shop, you can still use all
Microsoft products to create your SOAP application.
Since the Visual C++ support in the Microsoft SOAP Toolkit is essentially the same as the
Visual Basic support, I chose to forgo a Microsoft SOAP Toolkit example for Visual C++ in
this appendix. We’ll look at a third-party product called 4S4C (Simon’s Soap Server Services
for COM) instead. This is a product put out by Simon Fell and you can download it at
http://www.pocketsoap.com/4s4c/. I used the released 1.3.3 version for this book.

More developers create SOAP toolkits to meet specific needs every day. Not only do
third-party products provide support for SOAP on a variety of platforms; they also sup-
port many different languages as well. Trying to find a toolkit for a specific platform and
language could prove difficult given the diversity of the third-party product offerings.
However, the SoapWare.Org site at http://www.soapware.org/directory/4/
implementations can simplify the matter. While this list is less than complete, it does
provide the most complete list available on the Internet. You can search for your lan-
guage and platform of choice. In most cases, you’ll find at least one choice; in many
cases, you’ll find two or more possibilities. Another place to look for Visual C++ specific
solutions is Dave Angulo’s site at http://people.cs.uchicago.edu/~dangulo/
soap/. The entries on this site link directly to the vendor sites, so you get more infor-
mation than the SoapWare.Org site, but fewer listings. This site also contains a wealth of
information about SOAP in general, so it’s worthwhile visiting this site even if you’ve
already decided to use a specific SOAP toolkit.

The appendix begins with an overview of the 4S4C toolkit. Next, we’ll look at a typical appli-
cation for this toolkit. Finally, the appendix will show you how to create the application using
Visual C++ as the language of choice. This includes testing techniques and tips on handling
SOAP errors. By the time the appendix is finished, you’ll have a good idea of how Visual C++
development can differ from the Visual Basic development we discussed in the book.
 An Overview of the 4S4C SOAP Toolkit 391

An Overview of the 4S4C SOAP Toolkit

As previously mentioned, 4S4C is a third-party SOAP toolkit created by Simon Fell. Many
developers find this product superior to other offerings because it has special features not
found elsewhere. Of course, like every other SOAP product, 4S4C also has a few rough edges.
For example, the current version lacks a GUI-based or text-based WSDL generator. You gen-
erate the WSDL file using a Web page, which is usable, but not convenient. The following
sections will look at this product in detail and show you how to install it on your server.

Features
The first thing you’ll want to know about 4S4C is the set of features that it supports. One
of the most impressive aspects of the 4S4C SOAP Toolkit is the level of interoperability it
provides. Somehow the product author has managed to provide something the big names
haven’t so far, a way to allow at least the major platforms talk to each other most of the
time. This support includes both Microsoft SOAP Toolkit and Apache SOAP Toolkit
clients. You’ll find a list of other supported clients at http://www.pocketsoap.com/4s4c/
docs/interop.html. Of course, the author targeted 4S4C toward COM interoperability,
which means this product runs on a Windows server. It supports many client platforms, but
only one server platform, which is still a better deal than you’ll get from some vendors.

I chose 4S4C because it’s the most mature product available at the time of writing. By the
time you read this appendix, several other vendors will have mature (or at least complete)
products available and you should look at them all. For example, the IdooXoap offering
looks very interesting (http://www.zvon.org/ index.php?nav_id=34#c++) as
does the White Mesa product (http://www.whitemesa.com/
wmsoapsvc_about.htm). Make sure you consider all the available solutions on the
market. They all produce some level of SOAP output, but the capabilities vary by prod-
uct. The choice of product determines how quickly you produce an application and
how much code you’ll write to circumvent product shortcomings. Please write me at
[email protected] as you experiment with new products. I’ll try your suggestions as
I have time and post a review on my Web site at http://www.mwt.net/
~jmueller/ so others can benefit from your input.

The 4S4C SOAP Toolkit does rely on Internet Information Server (IIS). You can’t run it on
Windows 9x, but it works on both Windows NT and Windows 2000 setups. I agree with the
product author that running a major Web application on Windows 9x is doomed from the
start, so the lack of Windows 9x support isn’t a big deal.
Unlike the Microsoft SOAP Toolkit, 4S4C doesn’t need MSXML.DLL (or one of the App
newer versions such as MSXML3.DLL or MSXML4.DLL). This insulates it from changes
Microsoft may make to the DLL in the future. This also reduces some of the bulk you’ll find D
in other products. For example, the Microsoft SOAP Toolkit comes with the latest version of
MSXML.DLL, which increases the size of the download. (I have no less than three versions
of MSXML on my system right now, each of which have multiple files associated with it.)
392 Appendix D SOAP for Visual C++ Developers

You’ll note in Appendix C that XML doesn’t provide native support for many of the types
that you’ll probably use within a COM component. I’d love to say that 4S4C will solve all
of your woes, but it won’t. On the other hand, it does solve at least some of your woes. It
makes working with COM a lot less difficult by providing conversion for some COM types
that don’t ordinarily translate well. In addition, 4S4C also provides complete support for
complex or user defined types (UDTs). Simon is currently working on other conversions
and improvements of existing types as I write this. You can see the supported types list at
http://www.pocketsoap.com/4s4c/docs/datatypes.html.

4S4C supports the notion of a generic listener. You can create a single listener file that will
redirect queries to the appropriate component based upon the methodURI entry in the
SOAP message. We’ll see how this works later in the appendix. This is a welcome change
from the need to create a separate ASP for every application. The use of a configuration file
also makes management tasks easier. In short, this is one of those little ideas that have a big
impact on the developer.
Like the Microsoft SOAP Toolkit, 4S4C provides support for WSDL files. However, you
have a new option when working with 4S4C. Instead of creating a static WSDL file, you can
choose to create a dynamic WSDL file. 4S4C creates this file every time you call on the
server. This has the advantage of enabling you to update your application without worrying
about the WSDL file every time you make a change. The down side is that creating the
WSDL file dynamically wastes resources for an established application. You’ll still want to
create the WSDL file once the application development cycle is complete and you put the
application into production.
Many of the SOAP toolkit products currently on the market only support HTTL as a
transport. This limits them to real-time transmission of data where both client and server
must exist at the same time. Microsoft has already proven that distributed applications often
need asynchronous support where the client or server may be offline at the time of data
transmission. SOAP doesn’t support the complex infrastructure that technologies such as
Queued Components do, but it can support asynchronous communication using transports
such as Simple Mail Transfer Protocol (SMTP). 4S4C is one of the few products on the
market that support SMTP. Interestingly enough, 4S4C provides the SMTP service a script
(VBScript) that you can customize to meet your needs.
Like all SOAP toolkits, 4S4C supplies a dispatcher that does most of the work. The listener
grabs the data the client sends to it, transfers it to the dispatcher, which then instantiates
the correct component. 4S4C uses a dispatcher built on the SAX2 Parser
(http://www.megginson.com/SAX/index.html). It relies on Apache Xerces-C
(http://xml.apache.org/xerces-c/index.html) for parsing the request XML. The dispatcher
implements two interfaces: IDispatcher and IDispatcher2. Using these two interfaces, you
can create listeners that implement any suitable protocol. Theoretically, you can expand
4S4C to meet any need you might have. Learn more about creating your own listener at
http://www.pocketsoap.com/4s4c/docs/dispatcher.html.
 An Overview of the 4S4C SOAP Toolkit 393

Installation
4S4C uses the Microsoft Installer (MSI) file distribution method, so you’ll need to have MSI
file support on your server and workstation to use it. After you’ve downloaded 4S4C and
placed it on your server, you’ll want to install it. The following steps show you how.

Microsoft provides MSI support for most versions of Windows. Users of Windows 2000
and Windows ME have this support installed by default. You can download the required
version 1.2 support for Windows 9x at http://msdn.microsoft.com/code/
sample.asp?url=/ msdn-files/027/001/455/msdncompositedoc.xml and
Windows NT at http://msdn.microsoft.com/code/sample.asp?url=/
msdn-files/027/ 001/456/msdncompositedoc.xml. You can also create your
own MSI packages for distributing applications. The easiest method is using the Visual
Studio Installer found at http://msdn.microsoft.com/vstudio/downloads/
vsi11/default.asp. If your application has special installation needs, get the
Windows Installer SDK at http://msdn.microsoft.com/code/sample.
asp?url=/msdn-files/027/ 001/457/msdncompositedoc.xml or http://
msdn.microsoft.com/code/sample.asp?url=/msdn-files/027/001/530/
msdncompositedoc.xml.

1. Right-click 4S4C_<Version Number>.MSI and choose Install from the context menu.
You’ll see a 4S4C—InstallShield Wizard.
2. Click Next. You’ll see a License Agreement dialog.
3. Accept the license agreement terms after reading them. Click Next. You’ll see a
Customer Information dialog. This dialog allows you to install 4S4C for everyone
who uses the computer, or just for yourself.
4. Type the required user information. Choose an installation option. Click Next. You’ll
see a Setup Type dialog. In most cases, you’ll want to select the custom setup so that
you can control how 4S4C installs. You actually need to install the product twice: once
on the server and again on your workstation.
5. Select a Setup Type. The procedure assumes you’ll use a custom installation. You’ll see a
Custom Setup dialog like the one shown in Figure D.1. I installed the core components,
HTTP listener, SMTP listener, and Sample Servers and Clients on the server for my test
setup. I installed the samples and documentation on my workstation. You could place the
workstation files on the server’s hard drive in the suggested location. Using this approach
allows you to create a functional two-machine setup with everything installed in a mini-
mum of space. You’ll get an error message, however, when the installation program tries
to register the core components. I didn’t find that this created any problems.
App
6. Select custom setup options. Click Next. You’ll see a Ready to Install Program dialog.
7. Click Install. Depending on the options you chose, 4S4C will install the components
D
and files required to create SOAP applications using this toolkit.
394 Appendix D SOAP for Visual C++ Developers

Figure D.1
The 4S4C installation
provides several
installation choices.

Unlike the Microsoft SOAP Toolkit examples, I found the 4S4C examples worked as antici-
pated right out of the box. Part of the reason for the difference is that 4S4C dynamically
generates the WSDL files for you. This means it doesn’t assume anything about the location
of the client or the server. Everything is fresh every time you start the application. As previ-
ously mentioned, however, you’ll eventually want to create static WSDL files in the interest
of performance.

I have included a Visual C++ version of the Microsoft SOAP Toolkit interoperability
application (Demo.EXE) included with the 4S4C toolkit. You’ll find this sample, com-
plete with code, in the \Appendix D\Calc directory of the Que Web site for this book.
You can find it at www.quepublishing.com. The comments within the code com-
pletely document it. The C++ version uses the same access technique as the Visual
Basic version of the client application provided with the 4S4C toolkit. I also included
more error handling in this example than the example in this chapter. The example in
this chapter uses less error handling for the sake of explanation.

An Overview of the Application

The application for this appendix is a simple system information retrieval example. We’ll get
the computer name information for the server. You should already know the computer name
information for the remote test computer, so this is information that you can easily verify.
However, this example serves the larger role of showing how you can use SOAP to retrieve
remote computer information for administration purposes. Companies often use distributed
applications for administration and management purposes.
 Creating the Server-Side Component 395

We’ll write the server-side component in Visual C++ and expose it using 4S4C. Unlike the
Microsoft SOAP Toolkit examples in the book, we won’t generate a WSDL file and will
instead rely on 4S4C to generate it automatically. You could also call the component using
DCOM, if desired. In fact, you’ll find a DCOM version of the client in the Appendix D
directory of the Web site you can use for test purposes.
The client will use the generic RPC style found in the Microsoft SOAP Toolkit. Remember
that although we’re running the server side of the application using 4S4C, we’ll still need
another vendor product for the client side. However, you’ll find that this example works well
with any of the other clients supported by 4S4C. Interoperability with 4S4C is much better
than with the Microsoft SOAP Toolkit.

Creating the Server-Side Component

The server-side component in this example doesn’t do anything earth shattering, but it does
fulfill a potentially important service for network administrators. The component queries
the server and returns the server name, operating system type, version, build, and service
pack information (if any). In a real world environment, you’d probably ask this component
to do a lot more, but the fact remains that this is a practical service for today’s environment.
The following sections detail how you can build such a component. We’ll begin on a section
that tells how to create a static WSDL file using the 4S4C toolkit, since it doesn’t include a
utility for performing this task. The next section shows you the code you’ll use for the
example. There isn’t anything too special about this code, which is exactly the point. Using
4S4C doesn’t require you to change your current COM component code drastically. Finally,
we’ll talk about what you need to do to make the component usable with 4S4C. Many
developers find that it’s a lot less than they expected.

Generating a Static WSDL File

Earlier I mentioned that you had to create the WSDL file using a browser, rather than a
dedicated tool when working with 4S4C. At that point, you might have wondered why you
didn’t simply use the WSDL generation tool provided with the Microsoft SOAP Toolkit.
The answer lies in the interesting difference between these two products. Remember that a
single ASP page is all you need for a listener—that you won’t need to generate any other
pages for your applications as long as you can access this single page.
Also, remember that the magic behind this single listener approach is a CONFIG.XML file
located in the same directory as the SOAP.ASP file. The CONFIG.XML file contains a list
of all the components that you want to use and the GUID Windows uses to locate the
required interfaces in the registry. Put these two items together and you have a good idea
App
of why a Microsoft SOAP Toolkit generated WSDL file won’t work. Figure D.2 should
confirm your suspicions. D
396 Appendix D SOAP for Visual C++ Developers

Figure D.2
You’ll create a super
WSDL file when
using 4S4C.

As you can see, the WSDL that 4S4C generates works with all the components you have
registered for your Web site. This means that a single WSDL file will answer every client
request that relies on a WSDL file. Now, consider the implications of this difference from
the Microsoft SOAP Toolkit, which generates a unique WSDL file for each component.
The two toolkits aren’t compatible.
All you need to do after you see the WSDL file in the browser window, however, is use
Internet Explorer’s View/Source command to create a text file. Save this text file with the
same name as the ASP file (that is, SOAP.WSDL). You now have the static WSDL file
needed to reduce resource usage on the server. In some respects, this is easier than the
Microsoft approach. However, it’s also more error prone because you may find yourself
hand editing the file to overcome compatibility problems. Like most things SOAP, there
aren’t any simple or error-proof solutions to problems, just solutions that work better in a
given circumstance.

Developing the Server-Side Component Code

All I did for this example was create an ATL COM AppWizard project in Visual C++. Make
sure you choose a DLL project and add MFC support to it. I also allowed merging of proxy
and stub code.
Once the wizard created a component shell, I added a new ATL object (an ActiveX Server
component) to the project. I gave the new object the name ViewServer. The only attribute
that I set was to use the “Both” threading model. Make sure you disable any extra support.
For example, my version of Visual C++ includes an ASP tab because I installed the appropri-
ate Windows Platform SDK support. You want to disable this additional support because it
 Creating the Server-Side Component 397

complicates the example. The example code will still work, but you may end up with prob-
lems at the server end of the picture. In most cases, you’ll want to make your Visual C++
components as simple as possible.
Now that you have a component shell, it’s time to add some code to it. Listing D.1 contains
the code for the component. As you can see, there isn’t anything complicated about this
example, but it does form the basis of something useful. You could easily extend this compo-
nent to create an agent that you could use for remote server monitoring.

Listing D.1 Server Information Component

STDMETHODIMP CViewServer::GetServerInfo(BSTR *oResult)
{
AFX_MANAGE_STATE(AfxGetStaticModuleState())

CString oName; // Server Name.

OSVERSIONINFO VerInfo; // OS Version Information Stucture.
CString oMajor; // OS Major Version Number.
CString oMinor; // OS Minor Version Number.
CString oBuild; // OS Build Number.
CString oPlatform; // Type of OS.
CString oOutput; // Output String

// Get the Version Number.

VerInfo.dwOSVersionInfoSize = sizeof(OSVERSIONINFO);
GetVersionEx(&VerInfo);

// Determine the Platform Type

switch (VerInfo.dwPlatformId)
{
case VER_PLATFORM_WIN32s:
oPlatform = “Windows 3.1”;
break;
case VER_PLATFORM_WIN32_WINDOWS:
oPlatform = “Windows 9x or Windows ME”;
break;
case VER_PLATFORM_WIN32_NT:
oPlatform = “Windows NT/2000”;
break;
default:
oPlatform = “Unknown”;
}

// Convert the Build Infomration.

itoa(VerInfo.dwMajorVersion, oMajor.GetBuffer(5), 10);
oMajor.ReleaseBuffer(-1);
itoa(VerInfo.dwMinorVersion, oMinor.GetBuffer(5), 10);
oMinor.ReleaseBuffer(-1);
itoa(VerInfo.dwBuildNumber, oBuild.GetBuffer(5), 10);
App
oBuild.ReleaseBuffer(-1);

// Get the Computer Name.

D
ULONG ulBufferSize = MAX_COMPUTERNAME_LENGTH + 1;
GetComputerName(oName.GetBuffer(MAX_COMPUTERNAME_LENGTH + 1), &ulBufferSize);
oName.ReleaseBuffer(-1);
398 Appendix D SOAP for Visual C++ Developers

Listing D.1 Continued

// Create the Output String.
oOutput = “Computer Name: “ + oName +
“\n\rOS Name: “ + oPlatform +
“\n\rVersion: “ + oMajor + “.” + oMinor +
“\n\rBuild: “ + oBuild +
“\n\r” + VerInfo.szCSDVersion;

// Convert the Output String.

USES_CONVERSION;
*oResult = T2BSTR(oOutput.GetBuffer(512));
oOutput.ReleaseBuffer(-1);

return S_OK;
}

The only big issue with this code is the conversion of the return value. Notice that you need
to convert it to a BSTR. This is the best way to transfer data across the wire with SOAP.
Remember that SOAP won’t handle objects like CStrings very well. While you could
transfer a standard byte array, the BSTR is probably more efficient and interoperable.
I performed a local test of the component using a Visual Basic application. You’ll find this
application in the \Appendix D\Local Test directory of the Web site. Using Visual Basic
allowed me to check the performance of the component quickly, without wasting a lot of
time. Figure D.3 shows the dialog you should see when running this local test. Your dialog
may not match mine exactly, but it should contain the same type of information.

Figure D.3
Make sure you test
your component
locally to ensure it
works before placing
it on the server.

Understanding the Importance of Attributes

Some developers view the attributes within the IDL file of a component as part guesswork, part black magic.
While you can get by with less than accurate attributes in some environments, SOAP is very unforgiving in this
area. The methods used to create the SOAP message, the WSDL files, and other parts of the SOAP application
all require extremely accurate information from the component. This means that you can’t leave anything to
chance; you must use the right attributes in the right places.
Many of your applications will require some number of inputs to produce a single output. It’s important to mark
arguments for each method correctly. You’ll use the [in] attribute for those arguments that will only accept
values. The [in, out] attribute is useful for arguments that both accept and return values. Use the [out, retval]
attribute for the single return value of some methods. When you must return more than one value, mark all
outputs with the [in, out] or the [out] attribute. Here’s an example from the component in this appendix.
 Creating the Server-Side Component 399

interface IViewServer : IDispatch

{
[id(1), helpstring(“method GetServerInfo”)]
HRESULT GetServerInfo([out, retval]BSTR* oResult);
};

Note the placement of the attribute next to the single output of this component. You can also document the
attributes as part of the method declaration in the header file, but this isn’t required. The only place you must
carefully use attributes to define your component fully is within the IDL file.

Making the CONFIG.XML Entry

You need to perform two steps to make your component active on the server when using
4S4C. The first is to copy the component to some directory on the server and register it
using RegSvr32. When you register the component, you should see the standard success
message shown in Figure D.4. Note that an error here usually means there’s a significant
problem with your component, not that the component will actually work with the applica-
tion. Of course, I’m also assuming that you’re using Visual C++ 6.0 or creating an unman-
aged Visual C++ .NET component.

Figure D.4
An example of the
RegSvr32 success
message.

The second step is to create a new entry in the CONFIG.XML file for 4S4C. You’ll find
this file in the \Program Files\SimonFell\4S4C\HTTP\ASP of your server. When you ini-
tially open this file, you’ll see several other entries. These are the demonstration components
supplied with the products. You can leave these entries in place or remove them—it won’t
make any difference on the ability of the application to execute. However, keeping extra
entries in this file will reduce system performance, so removing the demonstration compo-
nents when you finish with them is a good idea. You’ll need to keep the opening and closing
<mappings> tags. The original CONFIG.XML file uses a name of demoService, but you can
use any service (mapping) name you like. Here are the entries you’ll make for this example.
(Note that the example <mapping> tag entry appears on several lines due to constraints in
the printed page width for this book; your entry should appear on one line.)
<mappings name=”demoService”>
<mapping name=”ServInfo” methodURI=”http://winserver/soapexamples/ServerInfo/”
progid=”ServInfo.ViewServer” iid=”{9D8E2FB6-FB7C-4542-9592-B2AB58CEF0FA}”/> App
</mappings>

At this point, you’ll need to stop IIS or ensure the CONFIG.XML file is unloaded in
D
some other way. IIS normally retains this file in memory for performance reasons. You can
verify the file is properly loaded into memory by directing your browser to the
http://<server name>/ssss4c/soap.asp?WSDL. You should see a display similar to the one
400 Appendix D SOAP for Visual C++ Developers

shown in Figure D.5. Note that I used a single <mapping> tag entry in this case for the sake of
clarity—your listing will be much longer if you retained the demonstration program mappings.

Figure D.5
Make sure you load
the mapping for your
component before
testing it with a client.

Creating the Client

Visual Basic is the easiest way to create a SOAP client because it offers direct support for the
components you’ll need to use. The Visual C++ version is harder because you need to work
with dispatch IDs. While this approach isn’t horribly difficult to use, you’ll find that there
are some situations where the complexity of using Visual C++ outweighs the advantages of
low-level access.
The following sections will show you how to create a basic dialog-based client application using
Visual C++. I’m assuming that you’ll use Visual C++ 6.0 and that you have all of the required
Microsoft SOAP Toolkit client support installed on your development workstation. While the
server-side component operates under 4S4C, the client side still requires Microsoft SOAP.
You’ll notice some differences because of the server-side setup, but the changes aren’t major.

Initial Setup
I purposely created a simple dialog-based application shell for this example to remove any
distractions. The project has a name of VC_SOAPTest, but you can use any name you like. I
set the dialog title entry on the MFC AppWizard—step 2 of 4 dialog to “Visual C++ SOAP
Test.” The About Box option is cleared. Figure D.6 shows the dialog setup. As you can see,
it’s quite simple. All you need to do is add a command button named IDC_TEST and pro-
vide a caption of Test.
 Creating the Client 401

Figure D.6
The example client
uses a very simple
dialog-box setup.

Adding the Code

You’ll need to add code to several files to make this example work. The first place you’ll
have to add code is in StdAfx.h, as shown here. Note the mssoap1.dll #import statement
must appear on a single line, although it appears on several lines in the book. Don’t try to
break this line because it will generate an error message during compiling.
// Begin special entries for SOAP.
#include “objbase.h”

#import “msxml3.dll”
using namespace MSXML2;

// Change path for your particular setup.

#import “H:\Program Files\Common Files\MSSoap\Binaries\mssoap1.dll”
➥exclude(“IStream”, “ISequentialStream”, “_LARGE_INTEGER”, “_ULARGE_INTEGER”,
➥“tagSTATSTG”, “_FILETIME”)
using namespace MSSOAPLib;

//End special entries for SOAP.

The two special entries are MSXML3.DLL and MSSOAP1.DLL. The MSXML3.DLL
provides the parsing mechanism for your SOAP message, while the MSSOAP1.DLL
contains the methods to interpret the SOAP message. As with every other example in the
book, you’ll need both of these DLLs to make the client work properly.
You’ll need to add an OnTest() for the Test pushbutton for the BN_CLICKED event. Listing App
D.2 contains the code for this event. Note that this example sticks to the SOAP essentials; I D
haven’t added much in the way of error handling. We’ll talk about error handling specifics
in the “Handling SOAP Errors” section of the appendix.
402 Appendix D SOAP for Visual C++ Developers

Listing D.2 OnTest Event Handler Code

void CVC_SOAPTestDlg::OnTest()
{
HRESULT hr; // Local result of method calls.
DISPID dispid; // Dispatch ID
DISPPARAMS dispparams; // Dispatch Parameters
VARIANT result; // Result of calculation.
CString ParamText; // Variable test from dialog.
EXCEPINFO ExceptInfo; // Exception information from SOAP
// invocation.
OLECHAR *pMethodName; // SOAP method to call.
ISOAPClientPtr m_pSoapClient; // SOAP Client Object

//Initialize the COM Library.

CoInitialize(NULL);

// Verify the SOAP client is NULL. After that, create a new instance of the
// SOAP client.
if (m_pSoapClient != NULL) m_pSoapClient.Release();
hr = m_pSoapClient.CreateInstance(__uuidof(SoapClient));
if(FAILED(hr))
MessageBox(“Cannot create SoapClient.”, “Error”, MB_OK |
MB_ICONEXCLAMATION);

// Make sure you include the WSDL path, service name, and port name. The WSDL
// path is specific to your Web server. The WSDL file contains the service
// name and port name. This information also appears in CONFIG.XML.
hr = m_pSoapClient->mssoapinit(_T(“http://WinServer/ssss4c/soap.asp?WSDL”),
_T(“demoService”),
_T(“ServInfoPort”),
_T(“”));
if(FAILED(hr))
MessageBox(“Cannot initialize SoapClient. “,
“Error”,
MB_OK | MB_ICONEXCLAMATION);

// Prepare the method variable.

pMethodName = L”GetServerInfo”;

// Create a mapping of the dispatch ID corrisponding to the method name.

hr = m_pSoapClient->GetIDsOfNames(IID_NULL,
&pMethodName,
1,
LOCALE_SYSTEM_DEFAULT, &dispid);
if(FAILED(hr))
{
MessageBox(“Cannot get dispatch id of calc method.”,
“Error”,
MB_OK | MB_ICONEXCLAMATION);
return;
}

// Initialize DISPPARAMS structure. No parameters are required, so set

// the number of arguments to 0.
dispparams.cArgs = 0;
dispparams.cNamedArgs = 0;
 Creating the Client 403

Listing D.2 CONTINUED

dispparams.rgdispidNamedArgs = NULL;

// Prepare result variant.

VariantInit(&result);

// Invoke the specified method.

hr = m_pSoapClient->Invoke(dispid,
IID_NULL,
LOCALE_SYSTEM_DEFAULT,
DISPATCH_METHOD,
&dispparams,
&result,
&ExceptInfo,
NULL);
if(FAILED(hr))
{
MessageBox(“Invoke of calc method failed.”,
“Error”,
MB_OK | MB_ICONEXCLAMATION);

}
else
{

// Display result.
ParamText = result.bstrVal;
MessageBox(ParamText, “Success”, MB_OK | MB_ICONINFORMATION);

// Clean up variants.
VariantClear(&result);
}

If you looked at the other examples in the book, you’ll notice the Visual C++ version is
much longer and more complex. However, it also provides you with precise control over the
client. The use of Visual C++ can change how the client interacts with the SOAP applica-
tion. For example, you have better control over the way output parameters are translated
into a form that SOAP can understand.
The client application works much the same as the Visual Basic examples. It begins by
instantiating a client object. The object uses mssoapinit() to create a connection with the
server. Notice that you must provide all four arguments when working with Visual C++.
Make certain that you set any unneeded parameters to a null string value as shown in the
example. Otherwise, SOAP will claim that it can’t create the connection for you.
However, once you create the connection, you’ll follow a different process than Visual Basic. App
The first step is to gain access to the dispatch ID of the component. You do this by specifying D
the name of the method you want to use. Notice that the first argument for the
GetIDsOfNames() method is set to IID_NULL. This value is currently reserved and you must
set the argument value to IID_NULL for every case. You can pass in either a single method
name or an array of names. The third argument sets the number of methods to return. If you
404 Appendix D SOAP for Visual C++ Developers

request an array of methods, then the dispatch ID variable will contain an array of dispatch IDs
on return. Otherwise, it contains a single dispatch ID that you can access directly.
This example doesn’t require the use of any arguments. If it had required parameters (as the
Calc example on the Web site), then you would have created an array of arguments as the
next step in the process. The array must be of type VARIANT and you must set the correct
variant type for each argument. Once you create this array, it’s referenced by the dispatch
parameters structure. Notice that the values in the dispparams structure for the example are
set to 0 because you have no arguments. Normally, you’ll set this value to a number that
corresponds to the number of entries in the argument array.
Calling the method comes next. You’ll use the client’s Invoke() method to perform this task.
Notice that you’ll need to pass the dispatch ID of the method that you want to call; not the
entire array of dispatch IDs. The three most important arguments are the dispatch parame-
ters structure, the result variable (another VARIANT), and the exception variable. On
return, the result variable contains the server information requested. Obviously, you’ll want
to clean up any variables you create once you retrieve the information from the server.
Now that you’ve looked at the application, it’s time to compile and test it. Figure D.7 shows
the results received from the example. Your results will differ depending on the name and
setup of your server.

Figure D.7
The server informa-
tion application pro-
vides details on the
name of the server
and the operating
system it uses.

Handling SOAP Errors

SOAP provides the same error handling capability using the client with Visual C++ as it
does for Visual Basic. The method used to access the data is slightly different, but will look
familiar to anyone who has worked with Visual C++ in the past.
As part of the task of learning about error handling for Visual C++, I decided to create a
clone of the Visual Basic example in Chapter 5, “Migrating an Application form DCOM to
SOAP.” The DisplaySOAPFault() method shown here provides similar functionality to that
example. (You can find the complete example in the \Appendix D\Throw Error directory of
the Web site for this book at www.quepublishing.com.)
void CThrowErrorDlg::DisplaySOAPFault(LPCTSTR pMessage)
{
 Handling SOAP Errors 405

HRESULT hr; // Result of call.

BSTR FaultString; // Fault string holder.
CString ErrorMsg; // Final error message.

// Begin building a message. Each method call will

// retrieve a different part of the fault string.
ErrorMsg = pMessage;
ErrorMsg = ErrorMsg + “\n\rFault Code: “;
hr = m_pSoapClient->get_faultcode(&FaultString);
ErrorMsg = ErrorMsg + FaultString;
ErrorMsg = ErrorMsg + “\n\rFault String: “;
hr = m_pSoapClient->get_faultstring(&FaultString);
ErrorMsg = ErrorMsg + FaultString;
ErrorMsg = ErrorMsg + “\n\rFault Actor: “;
hr = m_pSoapClient->get_faultactor(&FaultString);
ErrorMsg = ErrorMsg + FaultString;
ErrorMsg = ErrorMsg + “\n\rFault Detail: “;
hr = m_pSoapClient->get_detail(&FaultString);
ErrorMsg = ErrorMsg + FaultString;

// Free the string used to retrieve the data.

SysFreeString(FaultString);

// Display the error message.

MessageBox(ErrorMsg, “Error”, MB_OK | MB_ICONEXCLAMATION);
}

As you can see, many of the calls are the same as the Visual Basic example, but use syntax
that only a C++ programmer could love. The major concern is ensuring you free the fault
string once you’re finished collecting data. The example code relies on the SysFreeString()
call to perform the task. Figure D.8 shows the output from the ThrowError application.

Figure D.8
The ThrowError appli-
cation outputs the
same error informa-
tion as the example
in Chapter 5.

Notice that each call in the DisplaySOAPFault() method produces a result value. This actually
makes the Visual C++ detection scheme better than what you get with Visual Basic. You can
check the HRESULT after each call to ensure you still get valid information. Given the nature
of SOAP applications, the connection could break between calls. By detecting the HRESULT
value, you can provide the user with partial SOAP feedback, plus some additional information
about the cause of eventual connection failure. Here’s some code you could use to handle the
HRESULT values. App

LPVOID lpMsgBuf; // Message Buffer D

Cstring Msg; // Resulting Error Message
406 Appendix D SOAP for Visual C++ Developers

// Create the message.

FormatMessage(
FORMAT_MESSAGE_ALLOCATE_BUFFER |
FORMAT_MESSAGE_FROM_SYSTEM |
FORMAT_MESSAGE_IGNORE_INSERTS,
NULL,
hr,
MAKELANGID(LANG_NEUTRAL, SUBLANG_DEFAULT), // Default language
(LPTSTR) &lpMsgBuf,
0,
NULL
);

// Convert the message into the proper format and display it.
Msg = (LPCTSTR)lpMsgBuf;
MessageBox(Msg, “Error”, MB_OK | MB_ICONEXCLAMATION);

// Free the memory used by the message buffer.

LocalFree(lpMsgBuf);
return;

As you can see, all you need is a buffer and the HRESULT value to pass to the FormatMessage()
function. Since the return value is a void pointer, you’ll need to convert it into a string. After the
data is converted, you can display it as normal. The last step is to free the message buffer.
GLOSSARY
408 Glossary

This book includes a Glossary so that you can find terms and acronyms easily. It has several
important features of which you need to be aware. First, every acronym in the entire book is
listed here (even those you may already know). This way there is no doubt that you’ll be
able to find everything you need to use the book properly.
Second, these definitions are specific to the book. In other words, when you look through
this glossary, you’re seeing the words defined in the context in which the book uses them.
This might or might not always coincide with current industry usage because the computer
industry changes the meaning of words so often.
Finally, the definitions here use a conversational tone in most cases. This means they might
sacrifice a bit of puritanical accuracy for the sake of better understanding. The purpose of
this glossary is to define the terms in such a way that there’s little room for misunderstand-
ing the intent of the book as a whole.
Although this Glossary is a complete view of the words and acronyms in the book, you’ll run
into situations when you need to know more. No matter how closely I look at terms
throughout the book, it’s always possible I might miss the one acronym or term that you
really need to know. In addition, I’ve directed your attention to numerous online sources of
information. Few of the terms the Web site owners use will appear here unless I also chose
to use them in the book. Fortunately, many sites on the Internet provide partial or complete
Glossaries to fill in the gaps:
■ Acronym Finder (http://www.acronymfinder.com/)
■ Microsoft Encarta (http://encarta.msn.com/)
■ University of Texas Acronyms and Abbreviations (http://www-
hep.uta.edu/~variable/e_comm/pages/r_dic-en.htm)

■ Webopedia (http://webopedia.internet.com/)
■ yourDictionary.com (formerly A Web of Online Dictionaries) (http://www.
yourdictionary.com/)

Let’s talk about these Web sites a little more. Web sites normally provide acronyms or glos-
sary entries—not both. An acronym site only provides the definition for the acronym that
you want to learn about; it doesn’t provide an explanation of what the acronym means con-
cerning everyday computer use. The two extremes in this list are Acronym Finder
(acronyms only) and Webopedia (full-fledged glossary entries).
The owner of Acronym Finder doesn’t update the site as often as the University of Texas,
but Acronym Finder does have the advantage of providing an extremely large list of
acronyms from which to choose. At the time of this writing, the Acronym Finder sported
164,000 acronyms. The University of Texas site receives updates often and provides only
acronyms. (Another page at the same site includes a Glossary.)
Most of the Web sites that you’ll find for computer terms are free. In some cases, such as
Microsoft’s Encarta, you have to pay for the support provided. However, these locations are
still worth the effort because they ensure you understand the terms used in the jargon-filled
world of computing.
 Glossary 409

Webopedia has become one of my favorite places to visit because it provides encyclopedic
coverage of many computer terms and includes links to other Web sites. I like the fact that
if I don’t find a word I need, I can submit it to the Webopedia staff for addition to their dic-
tionary, making Webopedia a community-supported dictionary of the highest quality.
GL
One of the interesting features of the yourDictionary.com Web site is that it provides access
to more than one dictionary and in more than one language. If English isn’t your native
tongue, then this should be your Web site of choice. Now that we have the preliminaries
out of the way, it’s time to provide some definitions.
Active Directory Services Interface (ADSI) A set of APIs used to access Active
Directory, the central repository of information in Windows 2000. Active Directory is a hier-
archical database used to store many types of information in a somewhat freeform format.
ADSI allows access to both Active Directory data and the schema, which means you can use
it to create new database elements, as well as remove elements that are no longer in use.
Active Server Page (ASP) A special type of scripting language used by Windows NT
Server equipped with Internet Information Server (IIS). This specialized scripting language
allows the programmer to create very flexible Web server scripts. The use of variables and
other features, such as access to server variables, allows a programmer to create scripts that
can compensate for user and environmental needs as well as security concerns. ASP pages
use HTML to display content to the user.
Adaptive Network Security Alliance (ANSA) A proposed specification that ensures inter-
operability between security vendor application programming interfaces (APIs) using com-
mon interfaces for inspecting content. Competing standards include Open Platform for
Secure Enterprise Connectivity (OPSEC) Alliance and Common Content Inspection (CCI).
This product targets the Internet Security System (ISS) Group’s intrusion detection software.
Active Server Page (ASP) A special type of scripting language used by Windows servers
equipped with Internet Information Server (IIS). This specialized scripting language allows
the programmer to create flexible Web server scripts. The use of variables and other fea-
tures, such as access to server variables, allows a programmer to create scripts that can com-
pensate for user and environmental needs as well as security concerns. ASP uses HTML to
display content to the user.
ADSI See Active Directory Services Interface.
American Standard Code for Information Interchange (ASCII) A standard method of
equating the numeric representations available in a computer to human-readable form. The
number 32 represents a space, for example. The standard ASCII code contains 128 charac-
ters (7 bits). The extended ASCII code uses 8 bits for 256 characters. Display adapters from
the same machine type usually use the same upper 128 characters. Printers, however, might
reserve these upper 128 characters for nonstandard characters. Many Epson printers use
them for the italic representations of the lower 128 characters, however.
ANSA See Adaptive Network Security Alliance.
410 Glossary

Application The complete program or group of programs. An application is a complete

environment for performing one or more related tasks.
Application Service Provider (ASP) A vendor who provides downloadable or remotely
accessible service-oriented code using an Internet connection. Developers can use these
objects within their own code to obtain services from the ASP vendor. For example, an air-
line reservation system could rely on ASP modules from each of the major carriers to pro-
vide pricing and reservation support.
ASCII See American Standard Code for Information Interchange.
ASP See Active Server Page.
ASP See Application Service Provider.
AuthXML An XML-based technology that allows session authentication and authoriza-
tion using an Internet connection. Companies can also use this technology to exchange user
authentication lists and to verify the identity of remote users.
Bandwidth A measure of the amount of data a device can transfer in a given time.
Berkeley Internet Name Domain (BIND) This protocol allows Domain Name Service
(DNS) servers to translate human-readable Web site names to IP addresses. BIND was orig-
inally designed for use on UNIX systems.
Binary A method used to store worksheets and graphics files. Although you can use the
DOS TYPE command to send these files to the display, the contents of the file remain
unreadable. Other binary files include programs with extensions of EXE or COM.
Binary Compatibility Normally refers to two versions of an object that provide the same
interfaces at the lowest possible level. This term is also used generally to refer to two objects
of any type that provide low-level compatibility in some way. In some cases, the term refers
to two technologies and indicates the amount of low-level similarity the two technologies
provide.
Binary Large Object (BLOB) A special field in a database table that accepts objects such
as bitmaps, sounds, or text as input. This field is normally associated with the OLE capabili-
ties of a DBMS, but some third-party products make it possible to add BLOB support to
older database file formats, such as Xbase DBF file format. BLOB fields always imply OLE
client support by the DBMS.
BIND See Berkeley Internet Name Domain.
Biometrics A statistical method of scanning an individual’s unique characteristics, nor-
mally body parts, to ensure that an individual is who he says he is. Some of the scanned ele-
ments include voiceprints, irises, fingerprints, hands, and facial features. The two most
popular elements are irises and fingerprints because they’re the two that most people are
familiar with. The advantages of using biometrics are obvious. It not only prevents the user
from losing his identifying information (at least not very easily), but with proper scanning
techniques, it also protects the identifying information from being compromised.
 Glossary 411

BLOB See Binary Large Object.

Buffer The area in memory where program variable or other data is stored. For example,
applications will normally read more than one page from a word processed document to
improve performance. The applications store pages in addition to the one currently viewed GL
by the user in the buffer until needed.
Cascading Style Sheets (CSS) A method for defining a standard Web page template.
This might include headings, standard icons, backgrounds, and other features that would
tend to give each page at a particular Web site the same appearance. The reason for using
CSS includes speed of creating a Web site (it takes less time if you don’t have to create an
overall design for each page) and consistency. Changing the overall appearance of a Web site
also becomes as easy as changing the style sheet instead of each page individually.
CCI See Common Content Inspection.
CDATA See Character Data Section.
CDSA See Common Data Security Architecture.
Character Data (CDATA) Section Used to prevent interpretation of part of an XML
message by the XML parser. A CDATA section normally transports characters that the
XML parser would strip, mangle, modify, or simply change in undesirable ways. For exam-
ple, this is one way to get around the problems with transferring carriage return/linefeed
pairs in a message.
Character Set A general reference to a representation of printable or abstract symbols.
Character sets are normally encoded numeric forms of the symbology suitable for com-
puter storage.
Class ID (CLSID) A method of assigning a unique identifier to each object in the reg-
istry. Also refers to various high-level language constructs.
CLSID See Class ID.
COM See Component Object Model.
Common Content Inspection (CCI) A specification proposed by Aventail, Corp; Finjan
Software, Ltd.; and Check Point Software Technologies to ensure interoperability between
security vendor application programming interfaces (APIs). The CCI API seeks to promote
interoperability using common interfaces for inspecting content. Competing standards
include Adaptive Network Security Alliance (ANSA) and Open Platform for Secure
Enterprise Connectivity (OPSEC) Alliance.
Common Data Security Architecture (CDSA) A comprehensive set of security services
that will make secure transactions easier. It has a four-layer architecture: application, layered
services and middleware, Common Security Services Manager (CSSM) infrastructure, and
security service provider modules. CDSA is currently on The Open Group fast track to
becoming a standard.
412 Glossary

Common Object Request Broker Architecture (CORBA) The purpose of this proto-
col is to describe data and application code in a way that a variety of computer types can use.
It will eventually allow you to go to a Web page and download a mini-application (applet) as
part of that page. This is the Object Management Group’s (OMG) alternative to Microsoft’s
ActiveX. IBM originally designed CORBA for inclusion with OS/2, but other companies,
such as Sun Microsystems, now support this standard as well.
Component Object Model (COM) A Microsoft specification for an object-oriented
code and data encapsulation method and transference technique. It’s the basis for technolo-
gies such as OLE (object linking and embedding) and ActiveX (the replacement name for
OCXs, an object-oriented code library technology). COM is limited to local connections.
DCOM (distributed component object model) is the technology used to allow data transfers
and the use of OCXs within the Internet environment.
Connectivity A measure of the interactions between clients and servers. In many cases,
connectivity begins with the local machine and the interactions between applications and
components. Local area networks (LANs) introduce another level of connectivity with
machine-to-machine communications. Finally, wide area networks (WANs), metropolitan
area networks (MANs), intranets, and the Internet all introduce further levels of connectiv-
ity concerns.
Convent Vectoring Protocol (CVP) A proposed specification that ensures interoperabil-
ity between security vendor application programming interfaces (APIs) using common inter-
faces for inspecting content. Competing standards include Adaptive Network Security
Alliance (ANSA) and Common Content Inspection (CCI). This specification targets Check
Point’s Firewall-1.
CORBA See Common Object Request Broker Architecture.
Cracker A hacker (computer expert) who uses his skills for misdeeds on computer systems
where he has little or no authorized access. A cracker normally possesses specialty software
that allows easier access to the target network. In most cases, a cracker requires extensive
amounts of time to break the security for a system before he can enter it.
CSS See Cascading Style Sheets.
CVP See Convent Vectoring Protocol.
DASS See Distributed Authentication Security Service.
Data Conversion The act of changing data from one format to another. The success of a
data conversion depends on how well the converted data models the original data.
DCE See Distributed Computing Environment.
DCOM See Distributed Component Object Model.
Digital Signatures Initiative (DSI) A standard originated by the W3C (World Wide Web
Consortium) to overcome limitations of channel-level security. For example, channel-level
security can’t deal with documents and application semantics. A channel also doesn’t use the
 Glossary 413

Internet’s bandwidth very efficiently because all the processing takes place on the Internet
rather than the client or server. This standard defines a mathematical method for transferring
signatures—essentially a unique representation of a specific individual or company. DSI also
provides a new method for labeling security properties (PICS2) and a new format for asser-
GL
tions (PEP). This standard is also built on the PKCS #7 and X509.v3 standards.
DII See Dynamic Invocation Interface.
DIME See Direct Internet Message Encapsulation.
Direct Internet Message Encapsulation (DIME) An Internet media of the dime/
application type. It encapsulates multiple application-defined entities also known as payloads
into a single package. Each package can be of arbitrary size and type. The payload descrip-
tion includes data type, length, and an optional payload identifier.
DISCO See Discovery of Web Services.
Discovery of Web Services (DISCO) A service designed to make it easier to locate and
use SOAP services. This particular service is SOAP specific and a single vendor, Microsoft,
currently supports it. The DISCO service relies on a special protocol named SOAP Control
Language (SCL) to allow the discovery of services by remote computers.
Distributed Authentication Security Service (DASS) Defines an experimental method
for providing authentication services on the Internet. The goal of authentication, in this
case, is to verify who sent a message or request. Current password schemes have a number
of problems that DASS tries to solve. For example, it’s impossible to verify that the sender
of a password isn’t impersonating someone else. DASS provides authentication services in a
distributed environment. Distributed environments present special challenges because users
don’t log on to just one machine; they could conceivably log on to every machine on the
network.
Distributed Component Object Model (DCOM) The advanced form of the compo-
nent object model (COM) used by the Internet. This particular format enables data trans-
fers across the Internet or other non-local sources. It adds the capability to perform
asynchronous as well as synchronous data transfers, which prevents the client application
from becoming blocked as it waits for the server to respond. See Component Object Model
for more details.
Distributed Computing Environment (DCE) A specification created by the Open
Software Foundation (OSF) that defines methods for data exchange between a client and
server. The remote procedure call (RPC) support built into Windows NT is compatible
with the DCE specification.
Distributed System Object Model (DSOM) A full implementation of CORBA created
by IBM that fulfills the same purpose as Microsoft’s DCOM standard. DSOM is a binary
standard used on the network. The System Object Model (SOM) is the equivalent of COM
on the local machine.
DLL See Dynamic Link Library.
414 Glossary

DNS See Domain Name System.

Document Object Model (DOM) A method for describing the object representation
technique used within certain types of documents. Most people associate this term with
Internet-based documents such as those found on Web sites. The DOM determines how the
document presents objects such as links and text boxes.
Document Type Definition (DTD) A document that defines how an application
should interpret markup tags within an HTML, XML, or SGML document. In some
cases, such as HTML, the DTD is an actual specification. In other cases, such as XML,
the DTD is an external document supplied by the user or the vendor. A DTD can define
every characteristic of a document as long as those characteristics are defined using stan-
dard tags and attributes.
DOM See Document Object Model.
Domain Name System (DNS) An Internet technology that allows a user to refer to a
host computer by name rather than using its unique IP address.
DSI See Digital Signatures Initiative.
DSI See Dynamic Skeleton Interface.
DSOM See Distributed System Object Model.
DTD See Document Type Definition.
Dynamic Invocation Interface (DII) An interface that allows a client direct access to the
underlying request mechanisms for an ORB. This interface allows applications to dynami-
cally issue requests to objects without relying on the IDL interface-specific stubs. DII allows
blocking RPC-style requests, non-blocking synchronous requests, and send-only calls.
Dynamic Link Library (DLL) A specific form of application code loaded into memory
by request. It’s not executable by itself. A DLL does contain one or more discrete routines
that an application can use to provide specific features. For example, a DLL could provide a
common set of file dialog boxes used to access information on the hard drive. More than
one application can use the functions provided by a DLL, reducing overall memory require-
ments when more than one application is running.
Dynamic Skeleton Interface (DSI) An interface that allows a server direct access to the
underlying request mechanisms for an ORB. This interface allows servers to dynamically
respond to object requests without compile-time knowledge of the object implementation.
The client doesn’t know if it’s using DSI or RPC-style IDL Skeletons. DSI allows blocking
RPC-style requests, non-blocking synchronous requests, and send-only calls.
ebXML See Electronic Business eXtensible Markup Language.
Electronic Business eXtensible Markup Language (ebXML) A group of specifications
designed to standardize the use of XML globally. This effort will allow uniform business
communications based on XML and provide a consistent method to exchange data, create
relationships, and register business processes.
 Glossary 415

Encode The process of transforming a printable or abstract character into a coded format.
eXtensible Hypertext Markup Language (XHTML) A cross between XML and
HTML specifically designed for Net devices. Because this language relies on XML, most
developers classify it as an XML application builder. The language relies on several stan- GL
dardized namespaces to provide common data type and interface definitions. XHTML cre-
ates modules that are interpreted based on a specific platform’s requirements. This means
that a single document can serve the needs of many display devices.
eXtensible Markup Language Protocol (XMLP) An online communication protocol
alternative to SOAP. XMLP is still in the proposal stage, so there’s little information about
it. However, XMLP proponents state that it will provide a simplified method for transfer-
ring data while extending the capabilities of protocols, such as SOAP.
eXtensible Markup Language-Remote Procedure Call (XML-RPC) A predecessor to
SOAP that allows data exchange between two systems. XML-RPC relies on HTML type
organization within an XML framework. It provides some advanced features, such as com-
plex data type and array support. However, XML-RPC doesn’t provide support for such
crucial features as namespaces.
eXtensible Modeling Language (XML) A standardized Web page design language used
to incorporate data structuring within standard HTML documents. For example, you could
use XML to display database information using something other than forms or tables. It’s
actually a lightweight version of standardized generalized markup language (SGML) and is
supported by the SGML community. XML will also support tag extensions that will allow
various parts of a Web-based application to exchange information. For example, after a user
makes a choice within a catalog, that information can be added to an order entry form with
a minimum of effort on the part of the developer. Because XML is easy to extend, some
developers look at it as more of a base specification for other languages, rather than a com-
plete language.
eXtensible Rights Markup Language (XRML) A ContentGuard specification that
defines how a third party can use content provided by a host company. It describes the
rights, fees, and conditions of content. XRML also allows a vendor to define trusted systems
that can use a product for testing and evaluation purposes. This technology relies on a
trusted server to determine if someone can access content and what rights that person has
when he does.
FAQ See Frequently Asked Question.
Firewall A system designed to prevent unauthorized access to or from a network.
Firewalls are normally associated with Web sites connected to the Internet. A network
administrator can create a firewall using either hardware or software.
Frequently Asked Question (FAQ) A document that contains answers to questions that
many people ask. FAQs generally reduce support costs by providing answers to commonly
asked questions in one place. Vendors now use FAQs for many purposes, including both
hardware and software support.
416 Glossary

Globally Unique Identifier (GUID) A 128-bit number used to identify a component

object model (COM) object within the Windows registry. The GUID is used to find the
object definition and allow applications to create instances of that object. GUIDs can
include any type of object—even non-visual elements. In addition, some types of complex
objects are actually aggregates of simple objects. For example, an object that implements a
property page will normally have a minimum of two GUIDs: one for the property page and
another for the object.
GUID See Globally Unique Identifier.
Hacker An individual who works with computers at a low level, especially in the area of
security. A hacker normally possesses specialty software that allows easier access to the target
application or network. In most cases, hackers require extensive amounts of time to break the
security for a system before they can enter it. The two types of hackers include those that
break into systems for ethical purposes and those that do it to damage the system in some
way. The proper term for the second group is crackers. Some people have started to call the
first group “ethical hackers” to prevent confusion. Ethical hackers normally work for security
firms that specialize in finding holes in a company’s security. However, hackers work in a
wide range of computer arenas. For example, a person who writes low-level code (like that
found in a device driver) after reverse engineering an existing driver is technically a hacker.
HTTP See Hypertext Transfer Protocol.
HTTP Extension Framework ]See Hypertext Transfer Protocol Extension Framework.
Hypertext Transfer Protocol (HTTP) One of several common data transfer protocols
for the Internet. This particular protocol specializes in the display of onscreen information,
such as data entry forms or information displays. HTTP relies on HTML as a scripting lan-
guage for describing special screen display elements, although you can also use HTTP to
display non-formatted text.
Hypertext Transfer Protocol Extension Framework (HTTP Extension Framework)
Used with non-standard extensions to the HTTP header. The HTTP Extension
Framework describes which non-standard extensions a document contains. It also deter-
mines how the recipient should handle them. Common uses of an HTTP Extension
Framework include protocols such as SOAP.
IANA Internet Assigned Numbers Authority.
IDL See Interface Definition Language.
IETF See Internet Engineering Task Force.
IIOP See Internet Inter-ORB Protocol.
IIS See Internet Information Server.
Infrastructure The underlying base of an organization or system. One way to view infra-
structure is as the foundation on which all other elements of a system or organization are
attached. Many vendors use this term to indicate the compatibility of their product with
existing installations.
 Glossary 417

Interface Definition Language (IDL) A programming language construct used to define

the interfaces, methods, and parameters of a class. The IDL might use attributes to describe
some elements fully using a common methodology. In addition, the IDL normally includes
binary elements, such as interface identifiers. For example, COM relies on globally unique
GL
identifiers (GUIDs) for identification purposes.
Internet Engineering Task Force (IETF) The standards group is tasked with finding
solutions to pressing technology problems on the Internet. This group can approve stan-
dards created both within the organization itself and outside the organization as part of
other group efforts. For example, Microsoft has requested the approval of several new
Internet technologies through this group. If approved, the technologies would become an
Internet-wide standard performing data transfer and other specific kinds of tasks.
Internet Information Server (IIS) Microsoft’s full-fledged Web server that normally
runs under the Windows NT Server operating system. IIS includes all the features that you
would normally expect with a Web server: FTP, HTTP, and Gopher protocols along with
both mail and news services. Both Windows NT Workstation and Windows 95 can run
Personal Web Server (PWS), which is a scaled-down version of IIS.
Internet Inter-ORB Protocol (IIOP) A binary protocol that the Open Management
Group (OMG) designed for Internet use. This makes IIOP different from protocols such as
DCOM and CORBA that are designed for LAN use only. IIOP does perform better on the
Internet than DCOM or CORBA, but it has the same problems as DCOM and CORBA in
that it doesn’t communicate well through firewalls.
Internet Server Application Programming Interface (ISAPI) A set of function calls
and interface elements designed to make using Microsoft’s Internet Information Server (IIS)
and associated products such as Peer Web Server easier. Essentially, this set of API calls pro-
vides the programmer with access to the server. Such access makes it easier to provide full
server access to the Internet server through a series of ActiveX controls without the use of a
scripting language. ISAPI comes in two forms: filters and extensions. An extension replaces
current script-based technologies, such as CGI. Its main purpose is to provide dynamic
content to the user. A filter can extend the server by monitoring various events like user
requests for access in the background. You can use a filter to create various types of new
services, such as extended logging or specialized security schemes.
Interoperability A measure of an application’s ability to run in more than one environ-
ment, compatible or not. This term often refers to the ability of an application to run on
more than one operating system or hardware platform. In some cases, this term refers to
middleware’s ability to overcome interoperability problems between platforms.
ISAPI See Internet Server Application Programming Interface.
Java Document Object Model (JDOM) An object description technique that combines
the best features of the Simple API for XML (SAX) and the Document Object Model
(DOM). Few XML parsers currently support this new standard.
JDOM See Java Document Object Model.
418 Glossary

LAN See Local Area Network.

Local Area Network (LAN) Two or more devices connected together using a combina-
tion of hardware and software. The devices, normally computers and peripheral equipment
such as printers, are called nodes. An NIC (network interface card) provides the hardware
communication between nodes through an appropriate medium (cable or microwave trans-
mission.) There are two common types of LANs (also called networks). Peer-to-peer
networks allow each node to connect to any other node on the network with shareable
resources. This is a distributed method of files and peripheral devices. A client-server net-
work uses one or more servers to share resources. This is a centralized method of sharing
files and peripheral devices. A server provides resources to clients (usually workstations).
The most common server is the file server, which provides file-sharing resources. Other
server types include print servers and communication servers.
MAN See Metropolitan Area Network.
Mathematical Markup Language (MathML) An XML-based technique for describing
math notation. This includes both the structure and the content of the notation. It allows
standardized processing of complex equations over the Internet.
MathML See Mathematical Markup Language.
Message Transfer Agent (MTA) This is an X.400 standard term that refers to the part
of a message transfer system (MTS) responsible for interacting with the client. For example,
in an e-mail system, the MTA delivers e-mail to the individual users of that system.
Metropolitan Area Network (MAN) A partial extension and redefinition of the WAN, a
MAN connects two or more LANs together using a variety of methods. A MAN usually
encompasses more than one physical location within a limited geographical area, usually
within the same city or state. (A WAN can cover a larger geographical area, and sometimes
includes country-to-country communications.) Most MANs rely on microwave communica-
tions, fiber-optic connections, or leased telephone lines to provide the internetwork connec-
tions required to keep all nodes in the network talking with each other.
Microsoft Management Console (MMC) A special application that acts as an object
container for Windows management objects, such as Component Services and Computer
Management. The management objects are actually special components that provide inter-
faces that allow the user to access them within MMC to maintain and control the operation
of Windows. A developer can create special versions of these objects for application manage-
ment or other tasks. Using a single application like MMC helps maintain the same user
interface across all management applications.
Microsoft SOAP Messaging Object (SMO) A combination of component and wizard
that reduces the effort required to write SOAP-enabled components. It allows you to simu-
late the existing data exchange environment within the client. The client and server compo-
nent code remain about the same using this method because the major change occurs with
the listener and SMO code you need to write.
 Glossary 419

MIME See Multipurpose Internet Mail Extensions.

MMC See Microsoft Management Console.
MTA See Message Transfer Agent.
GL
Multipurpose Internet Mail Extensions (MIME) The standard method for defining the
content of Internet messages. This standard allows computers to exchange objects, character
sets, and multimedia using e-mail without regard to the computer’s underlying operating
system. MIME is defined in the IETF RFC1521 standard.
Namespace A method of organizing methods and other programming library resources
into easily accessible groups. Each group performs a given task or set of tasks on a particular
object type. For example, a namespace might contain several methods associated with the
file system on a computer. Although namespaces normally represent a means of organizing
methods, doing so is not required. Some developers use namespaces as a safe method for
placing methods in a container without regard to use.
Namespace Identifier (NID) A special number that identifies a particular company’s
resources. The NID is used with URNs to ensure uniqueness.
Network Interface Card (NIC) The device responsible for allowing a workstation to
communicate with the file server and other workstations. It provides the physical means for
creating the connection. The card plugs into an expansion slot in the computer. A cable that
attaches to the back of the card completes the communication path.
NIC See Network Interface Card.
NID See Namespace Identifier.
Object Request Broker (ORB) The component in CORBA that acts as middleware
between a client and server. The client makes a request of the ORB without knowing any-
thing about the server used to answer the request. Likewise, the server responds to the
request without knowing anything about the client.
OMG See Open Management Group.
Open Management Group (OMG) A consortium of more than 700 companies. The
goal of this consortium is to define better object-oriented programming methodologies.
OMG created the CORBA specification.
ORB See Object Request Broker.
OSF Open Software Foundation.
Parse To reduce a long label to its component parts. Spreadsheets normally break words
and numbers apart using the spaces between them as the break point. You can change how
spreadsheets parse a label by changing the format line.
PDA See Personal Digital Assistant.
420 Glossary

Peer-to-Peer Network A group of connected computers in which every computer can act
as a server and a client. Selected computers normally provide services to others, but unlike a
client/server network, the network administrator can distribute the processing load over sev-
eral machines. In addition, all nodes of a peer-to-peer network also act as workstations.
PEM1 See Privacy Enhanced Mail Part I.
Personal Digital Assistant (PDA) A very small PC normally used for personal tasks such
as taking notes and maintaining an itinerary during business trips. PDAs normally rely on
special operating systems and lack standard application support.
POST See Power-On Self Test.
Power-On Self Test (POST) The set of diagnostic and configuration routines that the
BIOS runs during system initialization. For example, the memory counter you see during
the boot sequence is part of this process.
Privacy Enhanced Mail Part I (PEM1) A specification that defines methods for encrypt-
ing mail in a way that protects the user’s identity but allows decrypting in the background.
This includes the use of keys and other forms of certificate management. Some of the speci-
fication is based on the CCITT X.400 standard.
Protocol A set of rules used to define a specific behavior. For example, protocols define how
networks transfer data. Think of a protocol as an ambassador who negotiates activities
between two countries. Without the ambassador, communication is difficult, if not impossible.
Proxy When used in the COM sense of the word, a proxy is the data structure that takes
the place of the application within the server’s address space. Any server responses to appli-
cation requests are passed to the proxy, marshaled by COM, and then passed to the applica-
tion.
RDDL See Resource Directory Description Language.
Remote Access The ability to use a remote resource as you would a local resource. In
some cases, this also means downloading the remote resource to use as a local resource.
Remote Method Invocation (RMI) A relatively simple wire protocol designed to sup-
port Java. RMI is a binary protocol like DCOM and CORBA and is built upon a CORBA
base. This protocol won’t support platforms other than Java.
Remote Procedure Call (RPC) One of several methods for accessing data within
another application. RPC is designed to look for the application first on the local worksta-
tion, and then across the network at the applications stored on other workstations. This is
an advanced capability that will eventually pave the way for decentralized applications.
Resource Directory Description Language (RDDL) This specification shows how an
organization could use a URL that it owns to point to an XHTML document that contains
a list of resources the company wants to make accessible. The document contains a descrip-
tion of the resource in human-readable form and embeds the required machine information
as part of the description.
 Glossary 421

Rivest Shamir Adleman algorithm (RSA) An authentication technology that relies on a

private-public key pair to create a set of credentials. The credentials are then used as a
means of identification for logging into various network resources. Using this methodology
allows for secure data transmission as well as user-oriented features, such as one password
GL
login to the network.
RMI See Remote Method Invocation.
RPC See Remote Procedure Call.
RSA See Rivest Shamir Adleman algorithm.
S/MIME See Secure/Multipurpose Internet Mail Extensions.
S2ML See Security Service Markup Language.
SAX See Simple API for XML.
Scalability A definition of an object’s ability to sustain increases in load. For example,
companies often rate networking systems by their ability to scale from one to many users.
Software scalability determines the ability of the software to run on more than one machine
when needed without making it appear that more than one machine is in use.
Scalable Vector Graphics (SVG) A vector-based method of describing a graphic using
XML. Vector graphics are infinitely scalable because they rely on math definitions, rather
than bitmaps. They also require less storage space than bitmap graphics. Unfortunately,
SVG requires more display time, processing power, and resources. They also require a
special XML parser and display application.
SCM See Service Control Manager.
Secure Hypertext Transfer Protocol (SHTTP) A technology designed to encrypt mes-
sages sent using the Internet. This technology is similar in purpose to Security Sockets
Layer (SSL). However, SSL secures the connection between two computers, while SHTTP
secures the individual messages. It’s possible to use both technologies together to provide
enhanced security.
Secure Socket Layer (SSL) A digital signature technology used for exchanging informa-
tion between a client and a server. Essentially, an SSL compliant server will request a digital
certificate from the client machine. The client can likewise request a digital certificate from
the server. Companies or individuals obtain these digital certificates from a third-party ven-
dors such as VeriSign, who can vouch for the identity of both parties.
Secure/Multipurpose Internet Mail Extensions (S/MIME) A secure method to transfer
attachments and other message elements on the Internet. S/MIME supports RSA’s public key
encryption technology. See Multipurpose Internet Mail Extensions for additional details.
Security Service Markup Language (S2ML) Provides a secure method for companies to
exchange information about transactions and customers. This protocol is commonly used in
business-to-business and business-to-customer environments. This technology relies on
XML as a basis for communicating data in a transparent manner.
422 Glossary

Service Control Manager (SCM) The SCM is part of the load balancing technology
used by Windows servers. When a client makes a DCOM call to the load-balancing router,
it’s the SCM that actually receives the request. The SCM looks up the component in the
load-balancing router table, then makes a DCOM call to one of the servers in the applica-
tion cluster to fulfill the request. The server in the application cluster creates an instance of
the request object, then passes the proxy for it directly to the client. At this point, the server
and the client are in direct communication; the router is no longer needed.
SGML See Standard Generalized Markup Language.
SHTTP See Secure Hypertext Transfer Protocol.
Simple API for XML (SAX) A less complex alternative to the Document Object Model
(DOM). This API is more efficient than DOM and will allow you to create faster applica-
tions. Unfortunately, SAX is read-only, so it limits the number of tasks you can perform.
SAX is also a developer-only specification; none of the standards groups such as W3C sup-
ports it.
Simple Object Access Protocol (SOAP) A Microsoft-sponsored protocol that provides
the means for exchanging data between COM and foreign component technologies like
Common Object Request Broker Architecture (CORBA) using XML as an intermediary.
Simple Object Access Protocol - Routing Protocol (SOAP-RP) A stateless protocol
used to exchange one-way SOAP messages between a client and server. The protocol sup-
ports intermediary destinations. You can also include a return path that enables two-way
communication between client and server. SOAP-RP requires a transport protocol such as
TCP, UDP, or HTTP.
Single Threaded Apartment (STA) A method of defining how object methods get exe-
cuted. STAs include three restrictions not found in multi-threaded apartments (MTAs). The
first is that an STA contains one, and only one, object. This ensures that once a component
is instantiated, that the resulting object doesn’t share memory space with any other object,
which could result in corruption. The second restriction is that one, and only one, thread
can enter the apartment to interact with the object inside. The reason for this restriction is
obvious. A single threaded object can only handle the requests of one thread at a time,
which means that COM must protect the object from access by more than one thread.
Ensuring that only one thread can enter the apartment at a time is the easiest way to accom-
plish this task. Finally, a thread can execute only one object method at a time. This restric-
tion ensures that there won’t be any data corruption due to shared variables within the
object. As a result of these restrictions, a single process could contain multiple STAs; one for
each STA object that the application instantiated.
Smart Card A type of user identification used in place of passwords. The use of a smart
card makes it much harder for a third party to break into a computer system using stolen
identification. However, a lost or stolen smart card still provides user access. The most
secure method of user identification is biometrics.
SMO See Microsoft SOAP Messaging Object.
 Glossary 423

SOAP See Simple Object Access Protocol.

SOAP-RP See Simple Object Access Protocol - Routing Protocol.
SOM See System Object Model.
GL
SSL See Secure Socket Layer.
STA See Single Threaded Apartment.
Standard Generalized Markup Language (SGML) A specification for defining docu-
ment format originally created for the publishing industry. Most developers consider SGML
too complex for standard display purposes. However, both XML and HTML are based on
SGML.
SVG See Scalable Vector Graphics.
System Object Model (SOM) An alternative object standard from IBM used with OS/2.
The Workplace Shell uses SOM in place of OLE to create objects.
TCP/IP See Transmission Control Protocol/Internet Protocol.
Transmission Control Protocol/Internet Protocol (TCP/IP) A standard communica-
tion line protocol developed by the United States Department of Defense. The protocol
defines how two devices talk to each other. Think of the protocol as a type of language used
by the two devices.
UDDI See Universal Description, Discovery, and Integration.
UDP See User Datagram Protocol.
Unicode Transformation Format (UTF) A standardized method of representing charac-
ters both printed and abstract using codes. Other forms of character representation include
ASCII.
Uniform Resource Identifier (URI) A generic term for all names and addresses that ref-
erence objects on the Internet. A URL is a specific type of URI. See Uniform Resource
Locator (URL).
Uniform Resource Locator (URL) A text representation of a specific location on the
Internet. URLs normally include the protocol (http:// for example), the target location
(world wide web or www), the domain or server name (mycompany), and a domain type (com
for commercial). It can also include a hierarchical location within that Web site. The URL
usually specifies a particular file on the Web server, although there are some situations where a
Web server will use a default filename. For example, asking the browser to find
http://www.mycompany.com, would probably display the DEFAULT.HTM file at that location.

Uniform Resource Name (URN) A managed resource identifier. Being a managed

resource means that a URN is guaranteed to be unique. Companies that want to use a URN
must apply for a Namespace Identifier (NID) from an authority such as the Internet
Assigned Numbers Authority (IANA). The NID appears as part of every resource reference
that the company creates.
424 Glossary

Universal Description, Discovery, and Integration (UDDI) A method of advertising

application and other software-related services online. The vendor offering the service regis-
ters at one or more centralized locations. Clients wanting to use the service add pointers to
the service to their application.
URI See Uniform Resource Identifier.
URL See Uniform Resource Locator.
URN See Uniform Resource Name.
User Datagram Protocol (UDP) Allows applications to exchange individual packets of
information over a TCP/IP network. UDP uses a combination of protocol ports and IP
addresses to get a message from one point of the network to another. More than one client
can use the same protocol port as long as all clients using the port have a unique IP address.
Protocol ports are of two types: well known and dynamically bound. The well-known port
assignments use the ports numbered between 1 and 255.
User Defined Type (UDT) A special construct supported by some programming lan-
guages that allows developers to create complex data types that closely mirror real-world
environments. Most developers include UDTs when a language-supplied simple or complex
type won’t fulfill a specific purpose.
UTF See Unicode Transformation Format.
W3C World Wide Web Consortium.
WAN See Wide Area Network.
Web Services Description Language (WSDL) A method for describing a service. The
file associated with this description contains the service description, port type, interface
description, individual method names, and parameter types. A WSDL relies on namespace
support to provide descriptions of common elements, such as data types. Most WSDL files
include references to two or more resources maintained by standards organizations to
ensure compatibility across implementations.
Wide Area Network (WAN) An extension of the local area network (LAN), a WAN
connects two or more LANs together using a variety of methods. A WAN usually encom-
passes more than one physical site, such as a building. Most WANs rely on microwave com-
munications, fiber-optic connections, or leased telephone lines to provide the internetwork
connections required to keep all nodes in the network talking with each other.
Wire Protocol A set of rules that govern the method for transferring data from one point to
another across a network. Wire protocols can transfer data in either binary or text format. In
addition, wire protocols can include additional features such as encryption and transactions.
WSDL See Web Services Description Language.
XDR See XML Data Reduced.
XHTML See eXtensible Hypertext Markup Language.
 Glossary 425

XLANG An automated process language developed by Microsoft. It is used to describe busi-

ness processes within products such as BizTalk server, but could be used for other purposes.
XML See eXtensible Modeling Language.
XML Data Reduced (XDR) A subset of the standard eXtensible Markup Language GL
(XML) schema. XDR is less complex than XML and therefore easier to implement. It con-
centrates on schema elements that the developers will most likely to use. Because XDR uses
a reduced schema, it doesn’t provide the flexibility found in XML.
XML Schema Definition (XSD) The portion of the XML specification that defines data
types and other data elements. It’s also related to a Web site containing such information by
use of XML parsers.
XMLP See eXtensible Markup Language Protocol.
XML-RPC See eXtensible Markup Language-Remote Procedure Call.
XRML See eXtensible Rights Markup Language.
XSD See XML Schema Definition.
 INDEX

Numbers
4S4C (Simons Soap Server
applications naming components, URIs and,
Services for Com), 54
client code, 106-109 126
toolkit, creating clients, 400-404
clients, creating with 4S4C output, surveys, 218-221
features, 391-392
toolkit, 400-404 partner access, 45
installing, 393-394
COM language binding, PDAs
overview, 391
136-137 Complex Type Example,
server-side components,
complexity differences, 98 testing, 331-332
creating, 395-400
components, registering, 399 Computer Name example,
Web site, 390
creating shortcuts, 100-102 testing, 338
WSDL files, creating, 395-396
data flow, 99-100 performance, 117-120
data type, testing, 254 remote access utilities, 163-165
A data viewers satellite, 46
creating client, 151-153 server-side code, 104-106
A Little Interface Definition separating data viewing logic server-side components, 104-105
Language (ALIDI), 136 from main component, creating with 4S4C toolkit,
access, 162 148-151 395-400
See also remote access server-side component, 147-148 testing, 109-116
Acronym Finder Web site, 127 updating, 145 databases, remote access,
databases, migrating, 153-155 272-274
acronyms, identifying (online debugging, 110-111 migrating to SOAP, 125
resources), 127 design basics, 99-100 updating for SOAP use, 243
ActiveX, Web-based differences and, 98-100 utilities
applications, 306-307 distributed compared to creating local components,
add-ons, PDAs, 324 client/server, 127 142-144
efficiency considerations, 124 migrating from DCOM, 140
AddIt client application form error handling, 404-406 updating client, 144-145
settings, 106-107 input, testing surveys, 214-216 updating server-side compo-
ALIDL (A Little Interface integrating modules, 134-136 nents, 141-142
Definition Language), 136 international, support for, 130 Web-based. See Web-based
analysis component, surveys, migrating applications
216-218 changing modules, 126-127 WSDL files, creating with
Apache, support for interna- error handling, 131-134 4S4C toolkit, 395-396
tional applications, 130 locating protocol problems, arguments, passing configuration
130-131 strings as, 167
Apache Toolkit, 83 performance concerns, 158-159
API (Application Programming prioritizing development, arrays, dynamic, databases and,
Interface) 128-129 262
Low-level, 86-87 problems with, 155 ASP files
functions protocols, 129-130 ISAPI Listener and, 179
GetComputerName( ), 293 reliability concerns, 156-157 scripts, paths, 37
GetComputerNameEx( ), security concerns, 157 Web-based applications,
292-293 troubleshooting, 159-160 316-317
OpenSCManager( ), 176 MIME and, 234-236 ATL COM AppWizard projects,
modules, testing, 134 creating, 396-398
428 attachments

attachments, 57-58 CDATA (character data), 204 complex data types, 245-254
MIME, 172 character substitution, data server-side component, 251-252
attributes transfer, 177 SOAP Toolkit and, 87
elements and, 118-119 UDTs, 248
ClearLogEntries( ) method, 188 WSDL generators, 247-248
importance of, 398-399
client code, applications, 106-109 Complex Type Example, PDAs
AuthXML, 68
client component, server status and, 329-332
viewer, 181-182
B client message type, 42
ComplexType 1.0 Type Library,
251
bandwidth, 26 client-side applications, data- ComplexType Component
bar-code scanners, PDAs, 323 bases, remote access, 267-272 Source Code, 252
Base64 coding, XML client-side components, 51 CompName component source
transmissions and, 205 Computer Name example, code, 293-294
PDAs, 333-335 component interactions, Web-
base64 data type, 347 creating data viewer, 151-153 based applications, 313-314
binary compatibility, 25 PDAs, Complex Type Example,
329-331 Component Object Model. See
binary data type, 347 COM
binary protocols, 25 client/server applications
compared to distributed, 127 components
BIND (Berkeley Internet Name limitations of, 146-147 access GUID, 53
Domain), 65 calling, 51
clients client-side, 51
binding, COM language binding, applications and, 99
136-137 server-side, 51
thick clients, Web-based
biometrics, 75-77 applications, 297-298 Computer Name example,
thin clients, 289-301 PDAs and, 332-338
BizTalk (Microsoft), 356-359
Editor utility, 359-362 form views, 298-299 configuration
Mapper utility, 363-365 thick clients and, 291 remote access and, 164
Orchestration Designer utility, Web page design, 299-301 strings, passing as argument, 167
365-368 updating, 144-145 connection loss, Web-based
reasons to use, 369-370 CLR (Common Language applications, 304
SOAP problems fixed, 368-369 Runtime), 102 connectivity, databases, 240
utilities, 359 CLSID (class ID), 18 controls
BLOB (binary large object) cmdCompleted_Click( ) data entry forms, 203
support, 257 method, 271 surveys, 203
block diagrams, planning migra- cmdMakeEntry_Click( ) converting
tion, 124 method, 190 See also migration
boolean data type, 347 cmdRemote_Click( ) source data types, 346, 353-354
Box, Don, 346 code, 181-182 converting data types, 245-247
Bray, Tim (XML), 198 cmdTask_Click( ) method, 271 CORBA (Common Objet
browsers, compatibility, Web- code Request Broker Architecture),
based applications and, 287 optimization, 119-120 240
toss-away, 172 IIOP, 19-20
bulletin boards, Web-based
interoperability and, 171
applications and, 286 color, PDAs, 340 SOAP and, 15, 19-20
byte data type, 347 COM (Component Object crackers, 45, 62, 64
Model)
CreateObject( ) function, 301
C GUIDs, 126
language binding, 136-137 cross-platform support, SOAP
case studies COM+ components, remote Toolkit, 88
reasons to use SOAP, 29-31 access and, 167 CSI (Computer Security
security, 78-79 Institute), security issues, 307
company locators, Web-based
SOAP solutions, 58-60
applications and, 286 CSS (cascading stylesheets)m
Web-based applications,
company policies, Web-based data entry forms, 233
317-319
applications and, 286
 document style WSDL files 429

D translation, 240 date data type, 347

XML, dateTime, 348 DCE (Distributed Computing
DASS (Distributed databases Environment), 15
Authentication Security arrays, dynamic arrays, 262 DCOM (Distributed
Service), 68 component simplicity, 256-257 Component Object Model),
data compression, databases connectivity, 240 126, 240
and, 239 data entry form application, 223 data transfer, 16-18
data type translations, 240 interoperability and, 171
data encryption, Masker 2.0, data types
374-376 migrating
complex, 245-254 database applications, 153-155
data entry validation, 244 from SOAP, 123
controls, 203 data viewers server-side components, 141-
formatting, XML and, 198 creating client, 151-153 142
round trips, 199-200 separating data viewing logic SOAP and, 15-18
data entry forms, 196-197 from main com-ponent, SOAP replacing, 46
application creation project, 148-151
server-side component, DCOM Wire Protocol, 15-18
221, 223-227
CDATA, 204 147-148 debuggers, scripts, Web-based
client application, 225-227 updating, 145 applications, 314-316
empty value processing, forms and, 242 debugging applications, 110-111
202-203 limitations of SOAP, 242-243
migrating, 153-155 decimal data type, 347
errors, 232-233
NULL value processing, multi-tier component, 263-266 declarations, complex data
202-203 order entry, 243 types, 349
performance, 230-231 parsing techniques, 244 design basics, applications,
privacy, 229-230 queries, 243 99-100
project, 236 remote access, 238-239
design tools, schemas as, 183
reliability, 231-232 client-side application, 267-272
interface, 249-251 detail, fault messages, 43
security, 228-229
shortcuts, 199-208 remote client creation, 252-253 developerWorks Web site, 123
templates, 233 server-side component, 251-252
development
third-party products, 206-208 shortcuts, 243-245
application modules,
vehicle choice, 197-199 testing application, 272-274
integrating, 134-136
WSDL files, 200-202 transactions, 278-279
IDE, enhancing efficiency of,
troubleshooting application,
data flow, 99-100 138
274-277
research tools, 137-138
data manipulation, remote uses/concerns, 239-243
access and, 163 reports, 243 DevelopMentor, 14
data security issues, 66-67 round trips, 199-200 DISCO (Discovery of Web
scalability, 240 Services), 48
data sharing, 12 security, remote access, 240 discovery protocols, 47
data transfer, 35, 52-57 server-side component, 255-262
formatting loss, 177-178 code generation, 259-263 display, PDAs, 338-341
HTTP and, 39-40 multiple, 258-259 distributed application use, 14
requirements, messages and, 39 SQLXML, 257-258 distributed applications
data transfers, DCOM, 16-18 SQL Server, defining, 255 client/server model and,
SQLXML, 257-258 146-147
data types, 346
state and, 154 compared to client/server, 127
application, testing, 254
survey forms, 210 security, 66
complex, 245-254, 349-351
surveys, 242
conversion, 245-247 Distributed Object Component
three-tier programming model,
conversions, 346, 353-354 Model. See DCOM
146-147
databases, validation, 244
toolkits, capacity, 244 DLLs (dynamic link libraries)
implementation, 351-352
transaction support, 240 namespaces and, 102-104
inline data typing, 352
transactions, remote access, SOAP Toolkit and, 84-85
method calls, migrating to
278-279 DNS (Domain Name Service), 65
SOAP, 124
troubleshooting, 279-281
overview, 346-349 document style WSDL files,
200, 202
430 DOM (document object model)

DOM (document object model), float data type, 348 GUID (globally unique
205 flowcharts, BizTalk Orchestratin identifier), 126, 250
compression and, 239 Designer, 366 components and, 53
XML parsing and, 205 GUIDGen, 250
foreign language code, encoding
DOS (denial of service) attacks, requirements, 130 GUIDs, GUIDGen and, 250
259
form views, thin clients, 298-299
double data type, 347
formats H
downloading native, 163
graphics, 163 StarOffice, 198 hackers, 62
SOAP toolkit, PDAs, 326-329 headers, 39
formatting
DSI (Digital Signatures data entry vehicles and, 197 processing, 170-171
Intiative), 68 data transfer, loss, 177-178 HTTP (Hypertext Transfer
DSOM (Distributed SOM), 19 data vehicles, XML and, 198 Protocol), 12, 22-23
dynamic arrays, databases and, forms data transfer and, 39-40
262 See also surveys databases, 240
data entry, 196-197 message portion, 37-38
E vehicle choice, 197-199
databases and, 242
HTTP Authentication
Framework, 71-73
eBay, 123 surveys, 208-221 HTTP Authentication
Web-based applications, 286 Framework RFC2617, 69
ebXML (Electronic Business
eXtensible Markup Language), functions HTTP Extension Frameworks,
47 API, OpenSCManager( ), 176 44-45
CreateObject( ), 301
EDI (Electronic Data EnumServicesStatus( ), 176 HTTP timeouts, PDAs, 332
Interchange), 357 EnumServiceStatus( ), 176 HTTP wrappers
elements, attributes and, 118-119 GetAllNames( ), 295-296 messages, 35
employee check-in application, GetComplexString, 251 response header portion, 38
remote access, 182-192 GetCompName( ), 295 human language support, Web-
GetComputerNameEx( ), based applications, 316
empty values
292-293
data entry forms, 202-203
ParseData( ), 268
surveys, 202-203
Replace( ), 182
I
encryption, 67
IBM Web Services Toolkit,
endBody( ) method, 103 G 311-312
EnumServicesStatus( ) function, IDE (integrated development
176 GetAllNames( ) function, 295-296
environment), enhancing
envelopes, messages, 22 GetComplexString method, 251 efficiency of, 138
error handling, 116-117, 131-134 GetCompName( ) function, 295 IDL files, method attributes,
Visual C++, 404-406 GetComputerName( ) API importance of, 398-399
Web-based applications, 303-307 function, 293 IdooXoap Web site, 391
error numbers, 133 GetComputerNameEx( ) API IdooXoap, PDAs, 328
errors function, 292-293
IETF (Internet Engineering
data entry forms, 232-233 GetObject( ) method, 187 Task Force), 68
surveys, 232-233 GetStatus( ) method, source IIOP (Internet Inter-ORB
code, 173-176 Protocol), 19-20
F GetTasks( ) method, 262 SOAP replacing, 46
fault messages, 41-44 GIF (Graphic Interface Format) IIS (Internet Information
categories, 42 images, 163 Server), paths, spaces in, 180
faultcode, fault messages, 43 graphics, downloading, 163 images, GIFs, 163
faultactor, fault messages, 43 GSS-API (Generic Security implementation
faultstring, fault messages, 43 Service Application Program data types, 351-352
file formats, native, 163 Interface), 69 problems, 53-55
 migrating 431

inline data typing, 352 check-in one component Message Validation Web site,
input applications, surveys, source code, 184-186 40-41
testing, 214-216 cmdRemote_Click( ) source messages, 35-37
code, 181-182 data transfer requirements, 39
installation Complex Type Client Code for
4S4C toolkit, 393-394 decoding, 54
PDA, 329-335 envelopes, 22
listeners, 51 CompNameProc source code, fault messages, 41-44
int data type, 348 295-296 categories, 42
integer data type, 348 Data exchange routines for header, 39
client application, 269-271 headers, POST action, 37
integrated development Data Survey Input client
environment. See IDE HTTP portion, 37-38
Source Code, 213-214 HTTP wrapper, response
integration, application modules, Data Survey Input Component header portion, 38
134-136 Source Code, 211-212 HTTP wrappers, 35
interfaces, databases, remote Data-Viewer Client Code, response header, error status, 38
access, 249-251 151-153 XML and, 35
GetStatus( ) method source XML portion, 38-39
international applications,
code, 174-176 XML wrappers, 35
support for, 130
IDL File Defining Visual Basic
Internet, security standards, Component Interface, 249-250 method calls, data types,
68-70 Movement routines for client migrating to SOAP, 124
Internet-only applications, 45 application, 272 methods
OnTest Event Handler Code, attributes, importance of,
interoperability, 171
402 398-399
online information about, 134
Remote Client Source Code, ClearLogEntries( ), 188
IPSec (Internet Protocol 252-253 cmdCompleted_Click( ), 271
Security Protocol), 69 remote test client source code, cmdMakeEntry_Click( ), 190
ISAPI Listener, 179-181 190-191 cmdTask_Click( ), 271
Server Information endBody( ), 103
Component, 397-398 GetComplexString, 251
J Server-Side Component for GetObject( ), 187
Java RMI (Remote Method SOAP Use, 148-151 GetStatus( ), source code,
Invocation), 163 Server-Side Database 173-176
SOAP and, 20-22 Component Accessible from GetTasks( ), 262
DCOM or SOAP, 260-262 mssoapinit( ), 108
Java RMIJScript, PDAs, 330 Thick Client source code, 297 SetComplete( ), 262
JVMs (Java Virtual Machines), 21 Utility routines for client- startBody( ), 103
application, 267-268 Microsoft BizTalk, 356-359
K Web Page Thin Client source
code, 300-301 Microsoft Installer Web site, 393
Kerberos Network live data Web-based application, Microsoft Queued Components,
Authentication Service, 70 302-303 35
kSOAP, PDAs, 328 local components, migrating, Microsoft SOAP Toolkit. See
142-144 SOAP Toolkit
L long data type, 348 migrating
applications
Low-level API, 86-87 performance concerns, 158-159
LANs (Local Area Networks), 12
problems with, 155
libraries M reliability concerns, 156-157
ComplexType 1.0 Type security concerns, 157
Library, 251 M-POST, 37 data types, 124
namespaces and, 102-104 data viewer applications, 145
maintenance, remote access and,
listeners, 86 164 creating client, 151-153
designing, 105-106 separating data viewing logic
installation, 51 Masker 2.0 data encryption tool, from main component,
374-376 148-151
listings
Analysis Component for Food memory, Web-based applica- server-side component, 147-148
Survey Source Code, 217-218 tions, 312-317 databases, 153-155
432 migration

migration N client-side component,

application modules Computer Name example,
changing, 126-127 Namespace Identifier (NID), 126 333-335
integrating, 134-136 color, 340
prioritizing development, namespaces, 22, 102-104 Complex Type Example,
128-129 WSDL, 49 329-332
block diagrams, 124 XML-RPC, 56 client-side component, 329-331
COM language binding, 136-137 naming, URIs and, 126 Computer Name example,
DCOM to SOAP, 123 native formats, 163 332-338
development research tools, development need, 323-324
137-138 networks, PDAs, 325 display issues, 338-341
distributed application concepts, NID (Namespace Identifier), 126 HTTP timeouts, 332
127-128 NULL values IdooXoap, 328
eBay and, 123 data entry forms, 202-203 implementation differences,
planning, 123-126 surveys, 202-203 Complex Type Example, 331
protocols, 129-130 JScript, 330
error handling, 131-134 kSOAP, 328
locating problems, 130-131 O modems, 323, 325
testing, 125 networks, 325
OMG (Object Management
troubleshooting, 142, 159-160 operating systems, 325-326
Group), CORBA and, 19
utility applications, 140 pocketSOAP, 327
creating local components, ONE (Open Network pointers, 340-341
142-144 Environment), 19 screen size issues, 339-340
updating client, 144-145 OpenOffice.org, 198 security issues, 341-343
updating server-side server-related differences,
OpenSCManager( ) API
components, 141-142 337-338
function, 176
VB projects, Project properties SOAP toolkit, downloading,
dialog box, 141 operating systems, PDAs, 326-329
WSDL files, trouble-shooting 325-326 SOAP::Lite Site, 326
applications, 138-140 order entry, databases, 243 special needs for, 323-326
MIME (Multipurpose Internet OSF (Open Software synchronization, 325
Mail Extensions), 57-58, 163 Foundation), 16 Trace Utility (MSSoapT),
applications and, 234-236 335-336
output applications, surveys, troubleshooting, 343-344
attachments, 172 218-221 Windows CE, 325-326
modems, PDAs, 323, 325
PDAs (Personal Digital
monikers, 18 P Assistants), 322-323
monitoring SOAP Toolkit and, 88
remote access and, 163 ParseData( ) function, 268
PEM1 (Privacy Enhanced Mail
security monitoring, 65 parsing techniques, databases, 244 Part 1), 69
monitoring programs, 46 partner access, applications, 45 performance
MSI (Microsoft Installer) Web passing configuration strings as applications, 117-120
site, 393 arguments, 167 issues, 26-27
mssoapinit( ) method, 108 paths, spaces (IIS), 180 surveys/data entry and, 230-231
MSSOAPR.DLL file, 85 payload protection, security, Personal Digital Assistants.
310-311 See PDAs
MSSoapT (Trace Utility), PDAs,
335-336 PCT (Private Communication phone book applications, remote
Technology), 69 access, 164-165
multi-tier component, databases,
263-266 PDAs platforms, application
add-ons, 324 differences, 98
Must Understand message type,
42 applications pocketSOAP, PDAs, 327
Complex Type Example, testing, pointers, PDAs, 340-341
MZTools (VB), 376-380 331-332
Web site, 138 Computer Name example, POST action portion,
testing, 338 message header, 37
bar-code scanners, 323
 security 433

privacy RegSvr, registering application RPC (Remote Procedure Call),

data entry forms, 229-230 components, 399 35, 86-87
issues, 65-66 reliability, surveys/data entry RPC style WSDL files, 200-202
surveys, 229-230 forms, 231-232
problem solving tools, 373-374 remote access S
problems solved by using SOAP, COM+ components and, 167
24-25 configuration and, 164 S/MIME (Secure Multipurpose
processing components, Web- data manipulation, 163 Internet Mail Extensions), 69,
based applications, thin clients, databases, 238-239 73-74
295-297 client-side application, 267-272 security issues, 309-310
interface, 249-251 S/WAN (Secure/Wide Area
productivity tools, 372-373 multi-tier component, 263-266 Network), 70
programming code, foreign lan- remote client creation, 252-253
guage encoding requirements, server-side component, 255-263 S2ML (Security Services
130 shortcuts, 243-245 Markup Language), 65, 70
projects, 120 testing application, 272-274 satellite applications, 46
data entry forms, 236 transactions, 278-279 scalability, databases, 240
surveys, 236 troubleshooting application,
schemas, as design tools, 183
274-277
protocols uses/concerns, 239-243 SCL (SOAP Control Language),
binary protocols, 25 maintenance, 164 48
discovery, 47 phone book application, 164-165 BizTalk and, 358
early computers and, 12 services and, 164
migrating applications, 129-130 SCM (service control manager),
streams and, 167 176
error handling, 131-134 task scheduling and, 164
locating problems, 130-131 screen size, PDA issues, 339-340
transport, 35 remote access utilities, 162,
scripting errors, Web-based
193-194
proxies applications, 304-305
applications, 163-165
thin clients, 290-291 components, existing, 168 scripts, debugging Web-based
VBWS Proxy Generator, employee check-in application, applications, 314-316
290-291 182-192 SDL (Service Description
proxy servers, 25 flexibility, 167 Langauge), 84
psWSDL Wizard, 200, 247-248, monitoring and, 163 security, 46, 62-63
381-383 non-issues, 170-172 application development,
overview, 162-172 integrating modules, 135
security issues, 169-170
Q server status viewer, 172-182
biometrics, 75-77
case study, 78-79
Web Services, 165-166 components, Web-based
QOP (quality of protection), 72
WSDL files and, 188-190 applications, 305
Que Web site , 141
remote client, database remote crackers, 45
queries, databases, 243 access, 252-253 data entry forms, 228-229
Queued Components Replace( ) function, 182 data security issues, 66-67
(Microsoft), 35 databases, remote access, 240
reports, databases, 243
distributed applications, 66
research tools, 137-138
R DOS attacks, 259
encryption, 67
Resource Directory Description
RDDL (Resource Directory Language. See RDDL error handling and, 116
Description Language), 126 HTTP Authentication
resources, Web-based applica-
Framework, 71-73
registration tions, 312-317
Internet standards, 68-70
application components, RMI (Remote Method issues, 64-65
RegSvr32, 399 Invocation), 15 migration considerations, 157
scripts, local components, 144 See also Java RMI monitoring, 65
registry entries ROAP.DLL, 84 payload protection, 310-311
avoiding ambiguity, 144 PDA issues, 341-343
round trips, data entry and
initialization process and, 167 remote access utilities, 169-170
surveys, 199-200
S/MIME, 73-74, 309-310
434 security

smart cards, 75-76 services, 47 Messaging Object Generator,

SSL, 74-75 DISCO, 48 94-97
standards, 67-75 remote access and, 164 overview, 83-98
surveys, 228-229 UDDI, 49-50 PDAs and, 88
two-way communications and, WSDL, 48-49 problems with, 87-89
310 Services MMC snap-in, 177 programming language support,
user identification issues, 75-77 390
user privacy issues, 65-66 ServicesDeclarations.BAS file, RPC (Remote Procedure Call),
vendors, 77-78 173 86-87
Web Services and, 165 SetComplete( ) method, 262 SMO (SOAP Message Object),
Web-based applications, 307-312 short data type, 348 86-87
WSML file, 180 variables, type information, 89
shortcuts Visual C++ and, 87
serializer, application creation, application creation, 100-102
101 WSDL Generator, 92-94
data entry forms, 199-200,
server message type, 42 202-208 SOAPAction entry, 37
server status viewer, 172-177, database remote access, 243-245 SOAP::Lite Site, 326
179-182 surveys, 199-200, 202-208 SOM (System Object Model), 19
client component, 181-182 SHTTP (Secure Hypertext source code
server-side component, 173, Transfer Protocol), 70 cmdRemote_Click( ) method,
176-177 smart cards, 75-76 181-182
Windows API and, 173 GetStatus( ) method, 173-176
SMO (SOAP Messaging
server-side code, applications, Object), 86-87 specification, problems with, 134
104-106 local components, creating, 143 speed, 27
server-side components, 51 migration from DCOM, 125
applications, 104-105 SQL Server, defining databases,
SOAP (Simple Object Access 255
data entry form application Protocol), 12
project, 223-225 contributors to, 14 SQLXML, 257-258
data viewers, 147-148 CORBA and, 15, 19-20 SSL (Secure Sockets Layer), 46,
database remote access, DCOM and, 15-18 70, 74-75
251-252, 255-262 HTTP and, 22-23 Web-based applications, 311
code generation, 259-263 implementation resources Web
multiple, 258-259 StarOffice, file formats, 198
site, 130
migrating, 141-142 Java RMI and, 20-22 startBody( ) method, 103
separating data viewing logic migrating from DCOM, 123 state, 128
from main component, overkill, 166-167 database applications and, 154
148-151 overview, 12-15 status, servers
server status viewer, 173, problems solved by using, 24-25 viewer, 172-182
176-177 specification problems, 134 working with information,
survey form, 210-212 state and, 128 177-179
Web-based applications, thin theory of, 34-35
clients, 292-295 XML and, 13, 22-23 streams, remote access and, 167
servers SOAP 1.1 Reference Web site, string data type, 348
application components, acti- 137 strings, configuration, passing as
vating with 4S4C, 399 argument, 167
PDAs, differences, 337-338 SOAP Messaging Object
Generator, 94-97 surveys, 196-197
proxy servers, 25
SOAP Messaging Object. See analysis component, 216-218
service information, polling, 173
SMO application processing, 221
status, working with
CDATA, 204
information, 177-179 SOAP Toolkit (Microsoft), 54-55 controls, 203
testing, 113-116 DLLs, 84-85 databases, 210
Web servers, 28 downloading for PDAs, 326-329 databases and, 242
service information, polling examples on two machines, 91 empty value processing, 202-203
servers for, 173 interoperability, 88 errors, 232-233
Service is in Use error, Web- limitations, 98 form creation, 208-221
based applications, 305 schemas, 153
Low-level API, 86-87
 varibles, type information, SOAP Tookit and, 435

forms toolkits, 53-54 UDTs (user defined types)

designing, 212-214 See also SOAP Toolkit complex data types, 248
service-side component, (Microsoft) SOAP Toolkit and, 87
210-212 4S4C uniform resource identifier
input application, testing, creating clients, 400-404 (URI), 37, 70, 126
214-216 creating server-side
NULL value processing, components, 395-400 uniform resource name (URN),
202-203 creating WSDL files, 395-396 126
output applications, designing, features, 391-392 unique identifiers, types of, 126
218-221 installing, 393-394 Universal Description,
output testing, 221 overview, 391 Discovery, and Integration
performance, 230-231 Web site, 390 specification. See UDDI
privacy, 229-230 capacity, databases, 244
project, 236 IdooXoap Web site, 391 updategrams, SQLXML, 258
reliability, 231-232 Web sites, 390 upgrades, SOAP Toolkit, 91
round trips, 199-200 White Mesa Web site, 391 URI (Uniform Resource
security, 228-229 tools Identifier), 37, 70, 126
shortcuts, 199-208, Masker 2.0, 374-376 URN (uniform resource name),
WSDL files, 200-202 MZTools (VB), 376-380 126
synchronization, PDAs, 325 problem solving, 373-374
user privacy issues, 65-66
productivity, 372-373
users, identification issues, 75-77
T psWSDL Wizard, 381-383
tcpTrace, 383-384 UTF (Unicode Transformation
task scheduling, remote access third-party, 372-388 Format), 38
and, 164 XML Spy, 384-388
utilities
tcpTrace, 383-384 toss-away code, 172 BizTalk Editor, 359-362
tcpTrace tool, 110-111 Trace Utility (MSSoapT), PDAs, BizTalk Mapper, 363-365
335-336 BizTalk Orchestration
telecommuters, 45 Designer, 365-368
tradeoffs in performance, 26
telephones, embedded SOAP migrating, 140
applications and, 341 transactions creating local components,
databases, 240 142-144
templates, data entry forms, 233 remote access, 278-279 updating client, 144-145
testing support for, 131 updating server-side
application modules, 134 transferring data, 52-57 components, 141-142
applications, migrating to remote access, 162, 193-194
SOAP, 125 translations, data types, 240
applications, 163-165
thick clients, Web-based transport protocols, 35 existing components and, 168
application design, 297-298 troubleshooting flexibility, 167
thick clients, using with thin, 291 applications, database remote non-issues, 170-172
access, 274-277 overview, 162-172
thin clients databases, 279-281 security, 169-170
form view clients, 298-299 migrating applications, 159-160 server status viewer, 172-182
processing components, 295-297 migration, 142 Web Services, 165-166
proxies, 290-291 migration problems, WSDL WSDL files and, 188-190
server-side components, files, 138-140 remote access employee check-
292-295 PDAs, 343-344 in application, 182-192
thick clients, using WSDL files, creating, 141
simultaneously, 291 tutorials, 123
WSDLGen, 247
Web page design, 299-301
Web-based applications, 289-301 U V
third-party tools, 372-388
UDDI (Universal Description,
three-tier programming model, Discovery, and Integration), validation, data types, databases
146-147 15, 49-50 and, 244
time data type, 348 BizTalk and, 357 variables, type information,
registry, 164 SOAP Toolkit and, 89
specification, 136
436 VB (Visual Basic)

VB (Visual Basic) toolkits, 390 document style, 200-202

IDE, enhancing, 138 tutorials, 123 input application testing, 214
migrating projects, Project White Mesa, 391 remote access utilities and,
Properties dialog box, 141 WSDL Generator, 141 188-190
MZTools, 376-380 Web-based applications, 284-285 RPC style, 200-202
SOAP Toolkit support, 390 ActiveX problems, 306-307 troubleshooting applications,
VBWS Proxy Generator, 290-291 ASP, 316-317 138-140
thin clients, 290-291 browser compatibility, 287 utilities for creating, 141
case study, 317-319 limitations of, 136
vendors, security, 77-78
challenges, 287-289 WSDL Generator
VeriSign, 74 component communication, complex data types, 247-248
VersionMismatch message type, 42 316-317 date variables, 263
viewers, server status, 172-182 component interactions, 313-314 SOAP Toolkit, 92-94
client component, 181-182 components, security, 305 SOAP Toolkit and, 90
server-side component, 173, connection loss, 304 Web site, 141
176-177 error handling, 303-307 WSDLGen utility, 247
human language support, 316
Visio, BizTalk and, 357 IBM Web Services Toolkit, WSML file, security, 180
Visual Basic. See VB 311-312
Visual C++ live data, 302-303 X
ATL COM AppWizard payload protection, 310-311
projects, creating, 396-398 processing component, thin XDR (XML Data Reduced) files,
error handling, 404-406 clients, 295-297 356
SOAP Toolkit and, 87 resource issues, 312-317 XHTML (eXtensible Hypertext
SOAP Toolkit support, 390 S/MIME, 309-310 Markup Language), 286
toolkit Web sites, 390 script debuggers, 314-316
XML (eXtensible Markup
scripting errors, 304-305
Language), 13, 22-23
security, 307-312
W server-side component, thin
See also ebXML
CDATA and, 204
clients, 292-295
WANs (Wide Area Networks), 12 document transmission
Service is in use error, 305
Web pages restrictions, 204-205
SSL, 311
design, thin clients, 299-301 formatting and, 198
thick clients, 297-298
dynamically created, 37 messages, 38-39
thin clients, 289-301
messages and, 35
Web servers, 28 form views, 298-299
SOAP and, 13
Web Services processing component, 295-297
Tim Bray, 198
remote access utilities, 165-166 server-side component, 292-295
wrappers, 35
security, 165 uses for, 285-287
XHTML, 286 XML Cover Pages, 50
Web Services Description XML Library, 85
Language. See WSDL WebServices Center Web site, 127
White Mesa Web site, 391 XML Parser 3.0 files, SOAP
Web sites Toolkit, 91
4S4C, 390 Windows API, server status
acronym identification, 127 viewer and, 173 XML Spy tool, 384-388
foreign language programming Windows CE, 325-326 XML Technology Protocol
encoding requirements, 130 Reference Web site, 127
Windows Internet Connector
IdooXoap, 391 XML-RPC, 52, 56
Library, 86
inter, 134
Message Validation, 40-41 Winer, Dave, 346 xmlns tag, 22
Microsoft Installer, 393 wizards, psWSDL Wizard, 200, XMLP (XML Protocol), 52, 56-57
MZ-Tools, 138 247-248 XRML (eXtensible Rights
Que, 141 Markup Language), 68
WSDL (Web Services Description
Resource Directory Description
Language), 15, 48-49, 352 XSD (XML Schema Definition),
Language, 126
BizTalk and, 358 55
SOAP 1.1 Reference, 137
files
SOAP implementation resources, XSI (XML Schema for Interfaces),
creating with 4S4C toolkit,
130 55
395-396
SOAP specification problems, 134