Four Macro Trends in Software Development Y2004

1
Four Macro Trends in Software Development Y2004

• Ivar Jacobson

Jaczone AB
ivar_at_jaczone.com
IBM Rational
2
Virtual Agenda

• Yesterday
• Today
• Tomorrow
• The best way to predict the future is to invent
  it!

Alan Kay
3
The 1980 Letter to The CEO of Ericsson

• Our component-based development approach will
  evolve into a world standard

• Proceed in three steps

• A standard approach including
• a modeling language like SDL, UML,
• a process like RUP supported by a first
  generation tool-set

• The modeling language becomes a formal, an
  abstract executable language like UML with
  action semantics

• Expert system support on top of process, language
  and tool talked about layman programming

4
Today

• We have the standard modeling language -- UML
• We have a standard development process -- RUP
• What next?

5
What is the Unified Modeling Language?

• The UML combines the best of the best from
• Use Case Modeling
• Data Modeling
• Business Modeling
• Object Modeling
• Component Modeling
• The UML is the standard language for modeling
  complex systems software systems, businesses
  business processes, combined hardware and
  software systems.

6
What is Rational Unified Process?

• The Rational Unified Process (RUP)is
  component-based
• It uses the Unified Modeling Language
• It is
• Iterative and controlled
• Use-case driven
• Architecture-centric
• Risk focused
• Configurable

7
Focus Moves to the Front-End
Now
Tomorrow

• Examples
• Use cases generate test cases and input to
  analysis
• Analysis will generate implementation eventually
  implementation will become superfluous for
  business software
• Test will be part of every activity test while
  you go forward

8
Macro Trends in support of the Front-End Move

• Use Cases and Aspects
• Executable UML
• Making Software Process Active
• Making Software Active (instead of Passive)

• Removed Trends (still very much trends)
• Quality from the Beginning
• Closing The Gap
• Domain-Specific Components Languages

9
Trend 1 Use Cases and Aspects

• Alternative trend names
• Keeping Use Cases Separate All The Way Down
• Aspect-Oriented Software Development

• Problems
• Adding new use cases/features to an existing
  system is painful
• Infrastructure extensions, for instance concerned
  with distribution, persistence, debugging,
  performance monitoring, auditing.
• Application extensions such as adding traffic
  recording to a base telecom system
• Separation of concerns (use cases) all the way
  down
• Over all models use case, analysis, design,
  implementation, executable code.
• Composition of concerns (use cases) again
• At some late point in time weaving back to an
  integrated whole

10
Today

• Consider a new feature for a base program
• Identify a location in the base program where you
  want to insert the new behavior
• Code the new behavior and insert it at the
  location in the base program
• The base program has been changed. The new
  feature was invasive to the base.

11
How would AOP help?

• AOP allows us to make the base program oblivious
  of the extensions ( aspects)

• AOP allows us to weave back extensions before
  execution
• UML has already support for extensions in use
  case modeling
• we need to extend UML to allow extensions between
  arbitrary design and implementation elements
• AOP allows us to keep use cases separate all the
  way

__________________________________________________
______________________________
ext p. 1
________________________________
ext p. 3
________________________________
Base
ext p. 2
________________________________
Extensions Aspects
Figure from Language Support for Changeable
Large Real Time Systems, Ivar Jacobson, 1986
12
Trend 2 Executable UML

• Fighting
• The Black Hole of Complexity
• in Software

The Two-Language Problem
13
UML Java or C

• Transitioning from UML to IDEs a Modern
  Approach
• Model Driven Architecture
• Frictionless design and Implementation
• Instant UML

IDE C or Java
Auto Round Trip Engineering
UML model files
Code files
14
The Two-Language Problem

• Two sets of files or programs
• The UML program and the C/Java code
• Two complex environments
• Existing IDEs are designed to be complete
• No assumptions on generating code from the UML
• Large overlaps between UML and IDEs
• IDEs only contribute 10-20 of the code, the rest
  is generated from the UML
• Huge investment in infrastructure in existing
  IDEs
• Compilers, debuggers, interpreters,
• Systemware, middleware, reusable assets in
  general

15
Making UML a formal language

• Actually a set of languages
• Specification language
• Programming language (that is executable UML)
• One program which combines visual and code like
  syntax

20
Expressiveness of Language
10
Smalltalk 80
Eiffel
Visual Basic
Java
C
Visual C
Ada
Simula
PERL
PL1
5
C
Pascal
COBOL
FORTRAN
Basic
Macro Assembly
1
Assembly
Language Evolution
1960s
1970s
1980s
1990s
2000s
16
Trend 3 Making Software Process Active

• Software development has never been
• as hard as today we need to SIMPLIFY
• Minimal training learn as you go
• Make process invisible yet very present
• Make it personal, make it light
• Without sacrificing quality
• Give them context-dependent, concrete advices
• Make people collaborate
• Let them focus on creative tasks
• A process engine in the hands of each developer
  and the whole team.

• Empower the developer

17
Use Next Generaton Technology Intelligent Agents

• Characteristics of intelligent agent
• Agents are proactive, adaptive and reactive
• Think of intelligent agents as objects driven by
  rules

18
Agents to assist the Software Developer

• Understand the software developers objectives
• milestones, roles, activities, artifacts
• Reason with relevant knowledge
• capture knowledge, and apply in context just in
  time
• Be a central part of a more complete and
  integrated development environment
• interact with other agents and productivity tools
• Be based on development standards
• development process and modeling RUP, UML,
• application platforms MS Web solution platform,
  J2EE,
• implementation technologies C, Java, XML,

19
Agents and Software Developers
20
An Example Creating a use case model
21
Agents on Top of RUP and Existing/New Tools
Project Mgr
Use Case Specifier
Software Architect
Use Case Designer
Component Designer
System Tester
A Layer of Agents
Requirment (use cases analysis)
Development (design impl)
Test
Project Mgmt
Content Mgmt
- - -
Change Management etc.
Reusable assets Application frameworks,
Middleware, Systemware
22
4 Making Software Active (instead of Passive)
Changing the World From Passive to Active Software
23
A Perspective of Automation Technologies
Benefit/Cost Ratio
Active Software
Software
Circuits
Year
1950
2000
1900
24
Our Approach Assistance in Context

• Empower
• the user

25
Contact information

• www.jaczone.com
• www.ivarjacobson.com
• Get my postcards at www.jaczone.com

26
Tomorrow, Life will be Much Better!
Summary

• We have UML, RUP and supporting tools
• Now focus will continue to move to the front-end
  of the software development cycle, driven by
• Use Cases and Aspects
• Executable UML
• Making Software Process Active
• Making Software Active (instead of Passive)
• Join the future

27
References

• Closing The Gap
• Closing The Gap Business Driven Enterprise
  Application Integration, Ivar Jacobson, RUC 2002
• RDA -- Ratonal Development Accelerators
  initiative (see http//www.rational.com/rda/ras/pr
  eview/index.htm).
• The Long Awaited Promise of Component Reuse is
  About to be Achieved, Ivar Jacobson Martin L
  Griss, Application Development Trends, June 2001
• Software Reuse Architecture, Process and
  Organization for Business Success (Addison
  Wesley) Ivar Jacobson, Martin Griss Patrik
  Jonsson, Addison Wesley Longman (1997)
• Use Cases and Aspects
• Language Support for Changeable Large Real Time
  Systems, Ivar Jacobson, Proceedings
  of OOPSLA86, pp 377-384, Sep 1986
• Use Cases and Aspects Working Together,
  Ivar Jacobson,
  soon to be published

28
References cont.d

• Executable UML
• Action Semantic for the UML, OMG ad/2001-03-01
• Executable UML, A Foundation for Model-Driven
  Architecture, Stephen Mellor Marc Balcer,
  Addison Wesley, 2002
• Making Software Process Active
• Call for Expert Systems, Ivar Jacobson Stefan
  Bylund, Application Development Trends, June
  2002, http//www.adtmag.com/article.asp?id3680
• A Multi-Agent System Assisting Software
  Developers, Ivar Jacobson Stefan Bylund,
  See www.jaczone.com/papers

29
Other Readings by Ivar Jacobson

• Object-Oriented Software Development--A Use Case
  Driven Approach (Addison Wesley)Jacobson et al,
  Addison Wesley Longman (1992)
• The Object Advantage Business Process
  Reengineering with Objects (Addison
  Wesley)Jacobson et al, Addison Wesley Longman
  (1994)
• Software Reuse Architecture, Process and
  Organization for Business Success (Addison
  Wesley) Ivar Jacobson, Martin Griss Patrik
  Jonsson, Addison Wesley Longman (1997)
• Unified Software Development ProcessJacobson,
  Booch, Rumbaugh, Addison Wesley Longman (1999)
• The Road to the Unified Software Development
  Process Ivar Jacobson, Stefan Bylund, Cambridge
  University Press, 2000