A WebML-Based Approach For The Development Of Web GIS Applications

B. Benatallah et al. (Eds.): WISE 2007, LNCS 4831, pp. 385–397, 2007.
© Springer-Verlag Berlin Heidelberg 2007
A WebML-Based Approach for the Development of
Web GIS Applications
S. Di Martino, F. Ferrucci, L. Paolino, M. Sebillo, G. Vitiello, and G. Avagliano
University of Salerno, Via Ponte Don Melillo, I-84084 Fisciano (SA), Italy
{sdimartino, fferrucci, lpaolino, msebillo, gvitiello}@unisa.it
Abstract. The goal of disseminating and manipulating spatial knowledge over
the Internet, has led to the development of Web applications, known as Web
Geographic Information Systems (Web GIS). Web GIS can be considered as a
particular class of data-intensive Web applications, since they are mainly
devoted to handle (spatial) information to and from the user. The success of
WebML (Web Modeling Language) for designing traditional data-intensive web
applications suggested use to extend this visual formalism to model relevant
interaction and navigation operations typical of Web GIS. In the paper, we
describe the proposed extension and provide an example of its application.
1 Introduction
eographic Information Systems (GIS) are a kind of software system specifically
suited to deal with data and attributes spatially referenced [9][10]. They turn out
to be fundamental in a great variety of contexts, ranging from urban planning to
marketing, route planning, natural disaster prevention, archaeology, and so on.
Recently, the growing need of sharing and exchanging geodata via Web has
affected researchers and professionals work, leading to the development of Web-
based GIS applications, known as Web GIS, intended for dissemination and
manipulation of spatial knowledge in specific domains. A Web GIS holds the
potential to make geographic information available to a worldwide audience, allowing
the Internet users to access GIS applications from their browsers without purchasing
proprietary GIS software. The requirement of GIS functionality via Web is also
proven by the success of Google Earth™
and Google Maps™
, which represent
examples of free solutions for navigating and querying thematic maps. The
development of Web GIS applications will be further boosted by the standardization
efforts made by the Open Geospatial Consortium, Inc.®
(OGC) [11].
With the increased time-to-market pressure, it is no longer possible to deal with
low-level issues, and create Web GIS applications from scratch. Thus, there is a
growing need for tools and methodologies that allow us to rapidly develop this kind
of Web applications and to rapidly modify them to meet the ever-changing business
needs. For that reason, leading companies, such as ESRI and Intergraph, have
provided users with proper extensions of their GIS products for developing Web GIS
applications [16][17][18][19]. GRASS and MapServer are well-established open
source GIS software which embed components for Web GIS implementation.
G

386 S. Di Martino et al.
However, such environments supporting the development of Web applications can be
very hard to use and require specialized skills [12].
For high-level development of traditional Web applications, many solutions are
currently available [1][3][7][8][15]. Among them, WebML (Web Modeling Language)
is a formal visual language specifically conceived to design data-intensive Web
applications [3][4][20]. It provides suitable models, a development process and a
CASE tool (WebRATIO) [5] to build applications that rely mainly on data
management and movements. WebML has been successfully applied in many
different contexts, from industry to academia, with positive results.
Web GIS can be considered as a particular class of data-intensive Web
applications, since they are mainly devoted to handle (spatial) information to and
from the user. Thus, in this paper we present an adaptation of the WebML approach,
to deal with the Web GIS context. The resulting methodology consists of two
orthogonal perspectives: Geodata and Metadata Conceptual Model and Hypertext
and Web Mapping Model. In particular, the Spatial E-R model by Calkins [2] has
been exploited to design the conceptual schema of (spatial) data. While, the WebML
Hypertext Model has been extended to include new notations specifically tailored for
GIS concepts and tasks. Moreover, the extension of the CASE tool WebRATIO will
provide potential Web GIS designers with a Rapid Application Development
environment, where they will be able to design in a visual fashion the application that
will be automatically generated by the tool.
The remainder of the paper is organized as follows: in Section 2 we recall the main
concepts of WebML, while in Section 3 we describe the proposed modeling concepts
and notations which extend WebML for Web GIS. In Section 4 we present an
example of use of the proposed notations, while the architecture of the WebGIS
applications generated by the WebRatio tool is described in Section 5. Some remarks
conclude the paper.
2 The WebML Methodology for Data-Intensive Web Applications
Graphical and diagrammatic representations play a central role in the field of software
and Web engineering [6]. They are widely employed to support many activities of the
software development process, such as specification, analysis and design, since they
help to describe and understand complex systems, by means of abstractions and
different views. WebML is a modeling language suited to support users in designing
data-intensive Web applications [4], by providing a set of visual notations to model
the content structure, data and navigational aspects of a Web application. Indeed, the
design of a Web application is based on two orthogonal perspectives: data and
navigation. The former is described by means of the well-known E-R data model, to
represent all the relevant entities and relationships. The latter is modeled by the
WebML Hypertext Model, which describes how the previously defined contents will
be arranged and rendered within the Web application. In particular, the application
structure and appearance are defined through four main levels of abstraction: site
views, areas, pages and content units.

A WebML-based Approach for the Development of Web GIS Applications 387
The site view is the highest level and is used to model the hypertextual structure of
the whole web site. A site view is composed of many areas that model the sections of
the application. Each area can contain both other areas (in a recursive way) and pages,
that are the containers of the information. Each page is composed of some content
units, the elementary pieces of information taken from the data schema. To bind a unit
with the corresponding data, the WebML Hypertext Model requires to specify a source
and a selector. The former is an entity of the database, the latter is a condition over
this entity, to filter the rendered data. These objects can be interconnected by links, to
model the resulting navigational aspects. Furthermore, links can be also used to
represent data movements, or parameter passing, between units. As an example, an
information gathered from the user via a web form, can be used in the selector
condition for extracting the data instances to be displayed. Several different types of
content units are defined in the WebML Hypertext Model. The well-defined semantics
and syntax of the WebML notation allowed for the construction of a specifically
conceived CASE tool, WebRatio, which automatically generates the Web application
starting from the corresponding design models, thus effectively reducing time and
costs of software development. Within this tool, each stereotype has both a visual and
an XML-based textual representation, and is used to specify additional detailed
properties, such as the grammar for the combination rules. The availability of an
XML specification has two main advantages: it enables the deployment of the same
design into multiple rendering formats, and it supports the definition of customized
extensions of the language.
3 An Extension of WebML for Web GIS
Web GIS are a special case of data-intensive Web applications, meant to deal with the
geographic data and share them across several users for different business goals. From
a design point of view, Web GIS present many specific characteristics, which make
them different from traditional data-intensive Web applications. As for the data model,
the complex nature of geographic data, where the two components, descriptive and
spatial, should be analyzed and managed in a joint manner, requires the use of a proper
modeling approach, named Geodata and Metadata Conceptual Model. In our
approach, we suggest the adoption of the Spatial E-R model [2], where each set of
geodata is described as a spatial entity characterized by a set of attributes, a geometry
and a couple of coordinates, whereas spatial associations among entities are expressed
in terms of topological relationships. In addition, since a Web GIS may also be aimed
at supporting data sharing, metadata management is a crucial issue to be accomplished.
Making interoperable data, which are heterogeneous in terms of formats and/or
sources, may guarantee better performances during data exchanging and retrieving.
As for the navigational model, specific functionality meant to dynamically navigate
a geographic map, commonly known as Web mapping, have to be taken into account,
giving rise to the Hypertext and Web Mapping Model as an extension of the WebML
Hypertext Model. In particular, to exploit WebML for modeling Web GIS, we have
considered new interaction metaphors able to capture the characteristics of this kind

of applications. Besides the common interactions a user performs when navigating
and executing a web application, we have taken into account the actions directly
performed on a map, when carrying out web mapping or spatial query tasks. Such
actions usually manipulate and produce data, which are not presented in a new web
page but are rather rendered onto the same map. Thus, new WebML units are needed
to model specific Web GIS interaction tasks, such as map visualization and
navigation. In the following we will describe some of them.
The MultiMap Unit
Basically, the MultiMap Unit is a visual metaphor for a map viewer, i.e. a graphic
component within a Web browser, able to render both vector and raster data arranged
in layers. It is characterized by the following properties:
• Name: the name chosen by the Application Architect for the MultiMap Unit.
• Raster Sources: the reference to raster entities. This source is optional, because a
Web GIS application may not require raster layers.
• Raster Selector [optional]: predicate useful to determine which objects
belonging to the raster entities share in the unit content.
• Vector Sources: the reference to vector entities. This source is optional, because
a Web GIS application may not require vector layers.
• Vector Selector [optional]: predicate useful to determine which objects from the
vector entities share the unit content.
Moreover, to correctly render source data, the MultiMap Unit requires a global
parameter, named Extension, which contains coordinates of the upper-left and lower-
right corners of the presented area. Each time this parameter is modified by an
external operation, the area rendered in the MultiMap Unit is updated accordingly.
(a) (b)
Fig. 1. Graphic notation of an instance of a MultiMap Unit (a) and its representation in a Web
browser (b)
Figure 1(a) depicts the visual notation of a MultiMap Unit instance, namely
TheMap, as well as the associated global parameter. The resulting representation in a
Web browser is shown in Figure 1(b). Here, both raster and vector layers are
displayed, as determined by invoking both source entities, while no selectors are set,

thus loading all the instances of the underlying data. Differently, when a Web GIS
application is expected to allow the activation and display of layers from a Table of
Contents on users’ demand, a different modeling is required as shown in Figure 2(a).
It contains a MultiMap Unit to render the layers, and two Multi-choice Index Units, to
represent the Table of Contents which is made up of two different legends. This
scenario also requires selectors associated with the MultiMap Unit sources, which are
used to select data in agreement with parameters set through the two instances of the
Multi-choice Index Units of WebML, namely RasterLegend and VectorLegend. The
resulting application on the Web browser has a checkbox, as shown in Figure 2(b), by
which users may activate a layer and visualize the corresponding data on the map.
Depending on aesthetic/usability factors, design layouts can also be provided where
legends are integrated into a homogeneous format.
(a) (b)
Fig. 2. Modeling the layer selection associated with a map
The Geometry Entry Unit
In a Web GIS, the spatial selection is one of the most common modalities of
interaction between the user and the application. It can be used either to select an area
of interest to magnify/reduce, or to select a set of geodata on which a basic
functionality may be applied. In order to model the associated geometry acquisition,
we have introduced the Geometry Entry Unit, which may be used to represent
interaction metaphors, typical of a Web GIS environment. This unit is characterized
by a name and the geometry type on which the interaction is based, which can be a:
• Point - it is a single point expressed as a coordinate (x,y). It corresponds to a
mouse click on the viewer map.
• Multipoint - it is a sequence of n points, corresponding to n mouse clicks on the
map.
• Line - it is constituted by two points, corresponding to 2 mouse clicks on the map.
The two points are then connected by a segment.
• Polyline – it is constituted by n points, corresponding to n mouse clicks on the
map. The points are connected by a sequence of consecutive lines, forming a single
polyline.

• Rectangle - it is made up of two points, corresponding to 2 mouse clicks on the
map, which are associated with the opposite vertexes of the rectangle diagonal.
• Polygon - it is constituted by n points, corresponding to n+1 mouse click on the
map (the first and last point are the same). It determines a polygon with n sides.
• Drag - it is constituted by 2 points, the former is captured when the user presses
the mouse button, the latter corresponds to its release.
Figure 3(a) shows the graphic notation of an instance of the Geometry Entry Unit,
which refers to a rectangle selection, namely the Rectangle Entry unit. The
corresponding Web representation is shown in Figure 3(b).
(a) (b)
Fig. 3. Graphic notation of a Geometry Entry Unit (a) and its representation in a Web browser (b)
The Zoom In/Out Units
These two units refer to the conventional zoom in and zoom out operations. Both take
three input values: the global parameter with the current extension of the map, the
global parameter with a magnification/diminishing factor and the instance Point of
Geometry Entry Unit, which is used to identify the area on which the operation should
be applied. As a result, if the operation succeeds, it sets a new value for the Extension
global parameter, according to the zoom factor. This arrangement is graphically
depicted in Figure 4 where an OK link leads to a new value for the Extension, while a
KO redirects towards a generic error page.
Fig. 4. An example of ZoomIn Unit

The Pan Unit
This unit represents the pan operation that shifts the focus of a map toward a specified
direction. This unit takes three input values, namely the two points of the map where
the user presses and releases the mouse button, and the global parameter with the
current extension of the map. If the operation succeeds, a new value for the Extension
global parameter is set, i.e. the new area of the map which is visualized. In Figure 5,
visual symbols involved in a pan operation are shown.
The CreateOverlay Unit
In a Web GIS application, a query processing usually outputs a set of geodata which
satisfy user’s criteria. To display such geodata onto a map, they are organized in a
new temporary layer, which may be overlaid onto the existing ones. Users may use
them to perform other basic operations and derive new geographic information. To
describe such a functionality through a WebML-like approach, we introduced the
CreateOverlay Unit, which can be exploited to model scenarios where users pose
queries and apply further spatial operations to the corresponding output in a cascade-
like fashion. The Unit requires two parameters, namely a spatial filter and a layer on
which the filter should be applied. In Fig. 9, an instance of the CreateOverlay unit is
depicted, namely NewTheme, where the spatial filter is obtained by a Geometry Entry
Unit which returns a rectangle, and the layer is the active one. In case the operation
succeeds, the OK link indicates that the OID of the newly created layer will be
forwarded to the MultiMap unit, which will eventually render it.
Fig. 5. An example of the Pan operation
Fig. 6. An example of the CreateOverlay operation

4 Modeling a Web GIS Application for Farm Houses Monitoring
The described web mapping units have been implemented in WebRatio, and adopted
in a preliminary prototype of the tool, to prove the effectiveness of the proposed
methodology. The prototype exploits a visual environment previously developed by
authors [13], which allows to design Spatial ER schemas and to generate a
corresponding relational database and an XML specification according to the
WebRatio format so that it can be used as underlying layer for the various units (both
new and old ones). In this section we describe the use of the tool in the design of an
application meant to monitor farm houses located in Campania, a region of Southern
Italy. Here, Health Districts may control phenomena related to distribution of livestock
and disease diffusion, by navigating a map, displaying geodata and posing queries
about areas of interest. Moreover, the resulting Web GIS application may support vets
and administrators in making decisions about policy and preventative plans.
Figure 7 depicts the framework for an interface layout for this kind of Web
application. It consists of a Map Working Area where geodata are displayed and users
may navigate, a Table of Contents, where a legend allows to select and activate
layers, a palette of GIS Tools to invoke GIS functionalities, a Query Frame where a
query may be posed, and a KeyMap which supports users in orienting themselves.
As for the design phase, in this section we report on two main aspects of the entire
process, namely:
• the description of geodata and their relationships by means of the Spatial E-R
extension, and
• the application of the proposed WebML extension for visually describing the
transitions among pages and the web mapping interaction tasks.
Fig. 7. A sketch of the expected WebGIS User Interface
As for the first step, Figure 8 depicts the conceptual schema which models entities
and relationships involved in this scenario. Three spatial entities, Region, Health
District and City, are used to describe the administrative boundaries of land boards in
Campania, while the FarmHouse spatial entity is provided to describe both farm house
properties and positions. City and Health District, as well as Health District and Region,
are connected by means of two aggregation relationships, which reflect the “part of”
association which exists between them. Analogously, farm houses are contained in at
least one city. Finally, information about vets is managed through a conventional entity.

In this case, a relationship between the Veterinary and Health District has been defined
in order to establish that a vet works just in one health district.
As for the Hypertext & Web Mapping design, the Web GIS site is composed by
three ANDed pages (see Fig. 12):
• a SideButton page that contains buttons to activate the GIS functionality, such as
Zoom In, Pan and SelectByLayer;
• a Central page, containing the ViewerGIS instance of the MultiMap Unit, the
Multi-choice units to de/select raster and vector layers within the map, an Index
Unit to select the active layer where the user computes an operation, and finally
some global parameters to store the active layer name, the map Extension and the
ZoomFactor;
• an Error Page.
Fig. 8. The Spatial E-R schema of the running example
In order to describe the zoom in functionality, we exploited the Zoom In Unit,
namely Magnify, which is fed through the ZoomFactor and the Extension global
parameters and an instance of a Geometry Entry Unit, which returns a point. The
result is a new map extension which will be stored into the Extension parameter. The
ViewerGIS Unit will automatically be refreshed.
For the Pan operation, the Move instance of the Pan Unit handles three parameters,
namely the present extension, the direction and the distance to move along that
direction. Also, the Drag instance of the Geometry Entry unit is modeled which
allows users to visually specify direction and distance by means of a rectangle drawn
by a drag operation on the map. Such values, together with coordinates received from
the Extension global parameter, are fed to the Move unit in order to compute the new
map extension. Finally, to visualize the new map boundaries we put the Move result
into the Extension parameter.
The last operation to model corresponds to a spatial query aimed to identify all the
farm houses within a user drawn rectangle. Then, in order to get graphical results, the
following steps are necessary:
• to add a new entry into the ViewerGIS source,
• an instance of the GeometricEntry Unit to draw the rectangle of the area of interest
on the map;
• the NewTheme instance of the CreateOverlay Unit in order to manage output
rendering.

Fig. 9. The WebML-based schema for the running example
Figure 10 shows the resulting Web GIS application, where, by following the same
approach, also other functions have been considered.
Fig. 10. The UI of the modeled Web GIS application
5 Model-Driven Generation of Web GIS Applications
Extendibility and flexibility are some of the key features characterizing WebRatio, the
CASE tool supporting the WebML methodology. Indeed, it provides a means to define

new symbols and links, and, for each of them, to specify the rules to generate the
corresponding J2EE-based code for the web applications. By exploiting this
mechanism, we have extended WebRatio to support the design of Web GIS
applications and their automatic generation starting from the specified Geodata and
Metadata Conceptual Model, and the Hypertext and Web Mapping Model. In
particular, the tool generates the server side code of the Web GIS application and the
corresponding database definition. The architecture of a generated Web GIS
application together with the employed technologies is depicted in Fig. 14.
The
Internet
Web Browser
HTTP Request
WFS/WMS
Postgres/
PostGIS
Geoserver
HTML/JAVAScript/
Raster
WebRatio
WebGIS Application
WebServer (Apache)
+ ApplicationServer
(Tomcat)
Generated
GIS
Application
Fig. 11. The client-server architecture of a generated Web GIS application
The application is based on a multi-tiered architecture and on a set of standard and
open technologies such as Apache/Tomcat, Postgres/Postgis, Geoserver, Web Feature
Service (WFS) and Web Map Service (WMS) protocols, which guarantee the
interoperability among server modules and the client browser. Basically, the WebML
units are translated into a JSP/Servlet program and successively mounted as a
Tomcat/Apache web application. This application will be able to manage requests
coming from the Internet and switch them to either the Geoserver or the PostGIS as
required. On the other side, the conceptual schema described through the Spatial ER
diagram is translated in an XML specification and input to WebRatio in order to build
the underlying database.
At run-time actions performed by the user on the client side are translated into
HTTP request by JavaScript code and sent to the server where data satisfying the
request are selected and sent back to the client as HTML, JavaScript and raster data.
In detail, a GIS query is transmitted as a GET or POST request written according to
either the WFS or WMS specifications. The request is therefore captured by the
generated GIS application and forwarded to Geoserver which, in turn, interprets the
query, composes the SQL statement according to the PostGIS DML and send it to the
DBMS. Once the DBMS computes the query, results are gathered by Geoserver to
create the answer. In particular, in case of a WMS request, Geoserver computes a
raster map containing the results encoded as a standard picture format (GIF, PNG,
SVG, etc.). In case of a WFS request, Geoserver collects data from the DBMS and

returns a GML (Geography Markup Language) encoded data to the generated GIS
server application. The latter further processes the resulting GML data, by sending it
back to the client side in HTML format
6 Conclusions
In the present paper, we described an approach for the design of Web GIS
applications based on WebML. In particular, we proposed to exploit the Spatial E-R
model to define spatial data and we introduced some visual notations to model
relevant interaction and navigation operations that are specific for Web GIS. These
new units have been implemented giving rise to an extension of the WebRatio tool for
the automatic generation of Web GIS.
As future work, we also plan to consider some other existing visual notations for
the specification of web applications (e.g. [1][14]) and study their extension to the
GIS domain. This would also allow us to perform a comparative analysis of different
approaches to the specification and generation of Web GIS applications.
References
[1] Baresi, L., Colazzo, S., Mainetti, L., Morasca, S.: W2000: A Modeling Notation for
Complex Web Applications. In: Mendes, E., Mosley, N. (eds.) Web Engineering, pp.
335–408. Springer, Heidelberg (2006)
[2] Calkins, H.W.: Entity Relationship Modeling of Spatial Data for Geographic
Information Systems. International Journal of Geographical Information Systems
(January 1996)
[3] Ceri, S., Fraternali, P., Bongio, A.: Web Modeling Language (WebML): a Modeling
Language for Designing Web Sites. Computer Networks: The International Journal of
Computer and Telecommunications Networking 33(1-6), 137–157 (2000)
[4] Ceri, S., Fraternali, P., Bongio, A., Brambilla, M., Comai, S., Matera, M.: Designing
Data-Intensive Web Applications. Morgan-Kaufmann Publishers, San Francisco (2002)
[5] Ceri, S., Fraternali, P., Matera, M.: WebML Application Frameworks, a Conceptual Tool
for Enhancing Design Reuse. In: Procs. ”Web Engineering”, Hong Kong (May 2001)
[6] Ferrucci, F., Tortora, G., Vitiello, G.: Exploiting Visual Languages in Software
Engineering - Handbook of Software Engineering and Knowledge Engineering (2001),
ISBN: 981-02-4973-X
[7] Garzotto, F., Paolini, P., Schwabe, D.: HDM- A Model-Based Approach to Hypertext
Application Design. ACM Transactions on Information Systems 11(1), 1–26 (1993)
[8] Isakowitz, T., Stohr, E.A., Balasubramanian, P.: RMM: A Design Methodology for
Structured Hypermedia Design. Communications of the ACM 38(8), 34–44 (1995)
[9] Longley, P.A., Goodchild, M.F., Maguire, D.J., Rhind, D.W.: Geographic Information
Systems and Science, p. 472. Wiley, Chichester (2001)
[10] Longley, P.A., Goodchild, M.F., Maguire, D.J., Rhind, D.W.: Geographical Information
System Principles, 2nd edn. Wiley, Chichester (2001)
[11] Open Geospatial Consortium, Inc. ® (OGC), http://www.opengeospatial.org

[12] Paolino, L., Sebillo, M., Tortora, G., Vitiello, G.: VIEW-GIS - An Environment
Supporting Web GIS Developers. In: VLC. Procs 12th International Conference on
Distributed Multimedia Systems - Workshop on Visual Languages and Computing,
Grand Canyon USA, pp. 141–146 (August 2006)
[13] Paolino, L., Sebillo, M., Tortora, G., Vitiello, G.: An OpenGIS® based approach to
define continuous field data within a visual environment. In: Bres, S., Laurini, R. (eds.)
VISUAL 2005. LNCS, vol. 3736, pp. 83–93. Springer, Heidelberg (2006)
[14] Pinet, F., Lbath, A.: A visual modelling language for distributed geographic information
systems. In: Proceedings IEEE International Symposium on Visual Languages
(September 10-13, 2000)
[15] Schwabe, D., Rossi, G.: An Object Oriented Approach to Web-based Applications
Design. Theory and Practice of Object Systems 4(4), 207–225 (1998)
[16] AutoDesk MapGuide, web site: http://usa.autodesk.com/adsk/servlet/index
[17] ESRI. ArcIms, web site: http://www.esri.com/software/arcgis/arcims/index.html
[18] Intergraph. WebMap, web site: http://imgs.intergraph.com/gmwm/
[19] Mapcell Web Site: http://www.territoriumonline.com/it/gis/mapaccel/
[20] WebML site. Available at http://www.Webml.org
View publication stats
View publication stats

A WebML-Based Approach For The Development Of Web GIS Applications

Recommended

More Related Content

Similar to A WebML-Based Approach For The Development Of Web GIS Applications (20)

More from Mary Montoya (20)

Recently uploaded (20)

A WebML-Based Approach For The Development Of Web GIS Applications