Software Engineering as the Next Level Up from Programming (Oracle Groundbreakers APAC Virtual Tour 2021, November 2021)

Software
Engineering
as the
Next Level Up
from
Programming
Oracle Groundbreakers APAC Virtual Tour 2021
Lucas Jellema, CTO & Architect AMIS | Conclusion

Programming  Software Engineering
Programming
“Writing code that works”
3

Programming  Software Engineering
Programming Software Engineering
“Writing code that works”
time
“all of the tools and processes an organization uses to build and
maintain that code over time. What practices can a software
organization introduce that will best keep its code valuable over the
long term? How can engineers make a codebase more sustainable
and the software engineering discipline itself more rigorous”
4

Your organization’s codebase
is sustainable when you are able to change all of
the things that you ought to change, safely, and
can do so for
the life of your codebase.
Software Engineering as the Next Level Up from Programming | APAC GB Tour 2021 5

Time and Change
• Functional Requirements because of new
insights, business evolution and innovation
• Non-Functional requirements because of
growth, costs and regulations
• Technological advances (& deprecations)
and shifts in community and among vendors
• Team members joining and leaving
• Changes in organization structure, scale,
processes
• Memories fading

Passage of Time
competitive advantage
breakthrough innovation
cutting edge
business critical
high visibility
buzzing with energy
sparkling new
7

Passage of Time
old
not modern
legacy
business critical
custom software
8

Passage of Time
end of scale
low on expertise
fragile
no tests, no specs, no docs
expensive TCO
high technical debt
no evolution
9

Not all code is created equal
• Demo in presentation or workshop
• Explorative / R&D / Doodles
• Assignment in study or training
• Snippet in a blog-article
• Proof of Concept
• Prototype
• MVP
• Alert Condition evaluation
• Product with a very short time to live (several weeks, not any longer)
• Contribution to open source project
• (internal) Reusable Library
• Operating system for guided nuclear missile/ SpaceX vehicle
• Unit-test for code component
• Healthcheck/smoketest
(potentially) Production : Professional Quality Software
Not Production Ready: not (necessarily) Professional Quality Software

What is software?

What is software?
Any instruction executed by a machine!

Types of code -
with distinctions in test, management, release , monitoring
Application
Platform
Infrastructure
Deploy/Provision
Time
Busines Run
Time
Tools/Macros/Scripts for
Design & Development
& Test & Process

Everything
is Code
15

Lifetime of a software application

Productivity as function of time
The Legacy Curve

Productivity as function of time and code quality
source: https://martinfowler.com/articles/is-quality-worth-cost.html

Working Software
• What work does it do?
• How can you tell?
Working Software

Small step for mankind…
FROM WORKING
SOFTWARE
TO PROFESSIONAL
SOFTWARE

Professional Software is ABLE Software
• Verifi
• Test
• Oper
• Read
• Evolv
• Maintain
• Observ
• Scal
• Recover
• Prov
• Afford
• Deploy
• Audit
• Impenetra
ABLE

Working => Professional Software
Working
Software
Professional
Software

Stages in Software Lifecyle
Functional Specification
Behavior & Unit test
Working Software
Professional Code
Built Software
Deployable Software =
DONE
Professional &
ABLE:
Live Software
(under Ops)
Developer
Team
Production
Preparation
Team
DoR
DoaD
DoD
PRoof
Pull Request
Peer Review
Production Ready
oPeRate
Put to Rest
25

Working
• according to functional specifications and technical interfaces
• proven
Working Software
26

Behavior Test
• The required
behavior as
experienced from
the outside
• specify
• document
• verify
Functional Specification
Behavior test
Working Software
Team
DoR

Unit Test
• Verify behavior of
• APIs & Interfaces
• Reusable elements
• Algorithms
• Aspects
• Functionality
• Non Functionality
• Happy & Non-Happy
• not: dependencies
28

Test is many things
Functional contract (specification and documentation)
Quick (REPL) feedback cycle for developer
Proof of “working”
Insulator that allows
refactoring and code optimizations
technical upgrades
Regression detector
for things changed
and things unchanged but impacted by changes
Health indicator & Smoke detector
Reference for (re)using code
29

• Who creates the test?
• and at what time?
• Who (or what?) executes the test
• at what moment | trigger?
• what is the outcome
• Who checks on the tests?
• specification coverage
• code coverage
• [real world] condition coverage
• timely execution of test and handling of
result

Definition
• functional specifications
clear, unambiguous and understood
• boundary conditions, non happy
cases and exceptional situations are
covered in specifications
• non-functional requirements are clear
• business value of feature is defined
• TCO budget is specified

Test Trophy
Live
32

Breach of Contract – at Dutch Court Systems
(no obvious technical debt)
Court Cases DB
Entity
Service
Composite
Business
Service
eCase
Portal
Team UI
Team Linking Pin
Team On Your Case
a check constraint
was added
users could no
longer submit
court cases
33

Breach of Contract – at Dutch Court Systems
Process Refinement
Court Cases DB
Entity
Service
Composite
Business
Service
eCase
Portal
Team UI
Team Linking Pin
Team On Your Case
Every team prepares test set
for other teams to automatically run
when they roll out any changes
34

QA: Static Code Analysis
• According to coding standards
• Consistent
• Readable (naming, formatting, length, …)
• Maintainable (Code Smells)
• Limited complexity
• Code Coverage (of tests)
• Vulnerabilities (in dependencies)
• Automatic improvements

Refactoring
36

Application Architecture
Business Logic
API/
Internal
Interface
Frontend
(web, chat,
public API)
Persistence
(of state) – pluggable
persistence providers
Internal &
External
Systems
Systems
Internal
Systems
External
Systems

Refactoring towards Clean Professional Code
• Compliance with
coding standards
• Reducing complexity
• Increasing readability
• Testable & test coverage
• Operable
• logging
• metrics
• configuration settings
• Life cycle management of technology stack &
technical debt
• Needed: Local build pipeline and runtime environment to quickly and frequently do
code analysis, pull & merge from master, build & automated test
Working Software
Professional Code Developer
DoaD
Unit tests & QA
(Behavior Tests)
Refactoring
38

Clean
Code
Guidelines
39

The Debt Mountain
Functional
Debt
Team Process Debt
Organization
Debt
Community
Debt
Ops Debt
Tech Debt
Product Owner
Architect
Team
Scrum Master

Technical Debt
• Hard coded styles in web pages
• Low (to no) test coverage
• Complex, deeply nested, large, hard to understand code units
• Meaningless or even confusing variable names
• High degree of coupling
• Low quality documentation
• Magic numbers (hard coded values) in program code
• Use of deprecated òr unsupported technologies
• Use of exotic features that no one understands but Jim
• Manual steps in CI/CD
• Frequent substantial redesign of
architecture/ platform / tech stack
• Libraries with security vulnerabilities
• Use of technologies not known or wanted by team
Ops Debt
Tech Debt
Impact/Risk (Longer Term)
• Changes increasingly become harder
(lengthy, costly, risky) – low agility
• Hard to keep/find & motivate technical
staff
• Drop in Team Productivity (velocity)
• Production incidents
• Increased Vulnerability
Owner: Architect (*
Found during: Dev,
Code QA, Review
41

What to do about debt?
• Identify debt during
• refinement, review, test, Ops intake, retrospective, audit, on boarding
• analysis, development, production usage
• Make debt explicit and visible – in a debt register and the risk log
• what and where
• severity, risk and impact (running cost!)
• resolution: how and effort
• Discuss debt (risks, running cost of not fixing) & plan actions
• in every sprint planning
• in every steering committee session
• …
• Continuously work on reducing debt – in small steps
• boy scout principle (improve everything you touch)
• explicitly set sprint budget aside
• Debt status should be a Team KPI
• Focus on a root cause: Team Process Debt
Functional
Debt
Team Process Debt
Ops Debt
Tech Debt
Organization
Debt
Community
Debt
Debt
Register
Risk
Log
Team
Story
Backlog
42

Prevent or at least Manage Tech Debt
• Design Principles & Reference Architecture
• Reusable Building Blocks
• Architecture Design Decisions
• identify/define
• process for deciding
• prototype
• record
• Application and Feature Portfolio Management
& Technical Life Cycle Management
• usage and state => risk
• effort to upgrade or replace
• Team Process:
• Definition of Ready – include solution design (how & non functionals)
• Definition of Done
• Peer Review
source: Oracle Cloud Reference Architecture
43

Team
• That story is in Janet’s area
• Sorry, Tom is on leave so we cannot work on X
right now
• Our tester is working on running all automated
tests
• Ellen is the only one on our team who can work on
the Python components
• Thomas knows how the CI/CD pipelines work
• I am not sure what business feature Sophie is
working on this sprint
• Bob built it, he knows how to demo it
• This [one year] old code is hard to maintain
because the person who built it has left the team

Metamorphosis – the miracle of the PR
45

PR
Pull Request
Peer Review
Production
Ready
Metamorphosis – the miracle of the PR

Pull Request == Please Review ?!
Team
DoaD
Appreciate my work
and learn from it
Help me improve it
and become a better developer
Take co-ownership of this code
47

Peer Review completes the Pull Request
• Peer Review completes (only) when
• Code is ABLE
• and beautiful
• the code is merged from the branch to the trunk
• and the tram may roll in
• because the peer considers the code their own
• Give priority to Peer Review!
• respond ASAP to Pull Request
• a proper Peer Review takes
real commitment and substantial time
(proportional to the complexity of the story)
Team
48

Impostor Syndrome
a psychological pattern in which an
individual doubts their skills, talents, or
accomplishments and has a
persistent internalized fear of being
exposed as a "fraud“
quite common among software developers
psychological safety in the team is crucial to
help members be free

Definition of Almost Done
• Code on trunk
• ABLE
• Compiles | Can be Built
• Satisfies
• QA
• Test (behavior & code)
• Non-functional characteristics (absolute & trend)
• Vulnerability
• Guidelines and Standards
• Automated CI/CD pipeline
• (covered by) Smoketest
• Technical Debt management
Professional Code
Built Software
50

Deployable == Done (as far as team is concerned)
• Deployable – but not yet deployed
• deploy decision is up to business
• CI/Continuous Delivery =
fully process up to deployability
• Continuous Deployment: automatic roll
out when DONE
Professional Code
Built Software
DONE
DoD
51

Deploy
• Business decision
• Automated – no touch
• Fusion of code & environment configuration
• Smoketest post deployment
(and periodically to check on health)
• Operations activated
Built Software
DONE
Professional &
ABLE:
Live Software
(under Ops)
Production
Preparation
52
Configuration

Canary
53

Observe the canary

Fix Forward
Fast Fix is not a special procedure nor a special environment
Tiered roll out
Spot issue [very
quickly]
Assign priority
Define test (to
reproduce and
guard against
regression)
Create fix on
branch
•working code
•professional code
PR and Merge to
trunk => DONE
<business
decision>
Roll Out
55

1202

60
Software Engineering as the Next Level Up from Programming
| APAC GB Tour 2021

Development
Production
scale,
performance,
security,
stability,
recoverability,
monitoring
61
Software Engineering as the Next Level Up from Programming
| APAC GB Tour 2021

Observing the Monkey
• “fire drills” – in live systems
• Testing Resilience of
• systems & automated procedures
• processes
• people
• By deliberately introducing errors in
runtime environments
• even in production ?!
• To test & prove & enhance measures

Professional Software – Generating Value for the Business

Software Wears Down

Life Cycle Management –
Functional and Technical Maintenance
• Ensure the up-date-ness of
the application and all its dependencies
• functional: which features are not used (and can be dropped), which are
suboptimal for users and require workarounds
• technical: which tech stack components are deprecated or even
forbidden, for which technology components do we no longer have the
skills?
• cost: which elements are way too
expensive for the value rendered?
Periodic Review
66

Life Cycle Management – Technical Maintenance
• A CEV vulnerability (CVE database https://www.cvedetails.com/)
• Release (or patch) of 3rd party library/framework
• New or deprecated (feature in) PaaS Service
• Custom pricing in used or unused service
• New specification from the business
• New non-functional requirement
• Incident/bug – functional or non-functional
• Technical debt assessment
• New version of platform component:
• e.g. Docker, Kubernetes, Java, Node
• New/custom architecture
choice/guideline
• New/custom coding standard
• New tool, new version of tool
• Law & Regulation, Ethical Insights
• Mere Progression of Time
Triggers for review (and change) – proactively monitored by the DevOps Team
Periodic Review
67

Automation
• Facilitate team members
• Reduce risk through frequent and small
steps
• Increase productivity of (scarce)
resources
• Reduce errors, increase quality
• Increase speed of process step execution
• Increase reaction speed (quicker value)
• Make things actually happen
• Scale up
• Reduce costs
source: https://enterprisersproject.com/article/2021/1/automation-4-trends-2021
68

Platform and Infra = Code
• Software Engineering practices apply to
definitions of Platform and Infrastructure
resources
• Agile team development process
• Source Code Control and Versioning
• Testing
• Code QA & Peer Review
• Automation
• Platform & Infra engineer are
software developers too

Take, Act On and Manage Design Decisions
• Any decision that has impact on technology stack,
team way of working, application wide requirements, toolset
• Made, shared and recorded in a structured and explicit way
• Record: What has been| is being | will be decided?
• label & title
• what, why & when
• current status (could be superseded)
• context: place in bigger picture, relation with other decisions
• criteria – based on what will we come to a decision
• options to choose from
• decision – what was decided
• why? why not the alternatives?
• who & when
• consequences/impact/actions
• if and when should we evaluate?
source: Fundamentals of Software
Architecture
,Mark Richards, Neal Ford 70

Shift Left
• Start attending to important aspects of the code base as early as possible
(and strive for design and prevention i/o detection & fix)
• security
• testing
• pair programming or
early peer review
• trunk based dev
• team ownership
• build & deploy
• operations
• automation
• TCO
Source: van der Cruijsen 2017
71

Thank you
for your attention
I hope
this was
useful
lucas.jellema@amis.nl | technology.amis.nl | @lucasjellema | lucas-jellema
72

Software Engineering as the Next Level Up from Programming (Oracle Groundbreakers APAC Virtual Tour 2021, November 2021)

Recommended

More Related Content

What's hot (20)

Similar to Software Engineering as the Next Level Up from Programming (Oracle Groundbreakers APAC Virtual Tour 2021, November 2021) (20)

More from Lucas Jellema (20)

Recently uploaded (20)

Software Engineering as the Next Level Up from Programming (Oracle Groundbreakers APAC Virtual Tour 2021, November 2021)

Editor's Notes