Working Group Summaries

1
Working Group Summaries

• WICSA Wrap-Up

2
Impression
We need more its not good enough yet
This is what architects really do help us here
too
3
General observations

• WICSA ? WCRE interaction worked
• Were stretching the bounds of software
  architecture
• Architects are senior technical leaders
• Architecture of the software part is hard to
  separate from architecture of the system
• Participation from practicing architects has led
  to outstanding discussions
• As a researcher, I really, really appreciate this
• Please tell the organizers how to extend this

4
Working Groups

• Education
• Hans van Vliet and Mary Shaw
• Quality
• Len Bass
• Architecting
• Christine Hofmeister
• Evaluation Analysis
• David Garlan

• ADLs in Practice
• Eóin Woods
• Documentation in Practice
• Paul Clements
• Components and Services
• Ivica Crnkovic
• Architectural Design Decisions
• Jeff Tyree

5
WG Education
Level? Topic ? CS undergr new programmer new architect
people skills 0 3 3-5
business 0 1 3
arch techn 1 2 4
reqs eng 1 1 5
programming 2 4 2
basic SE 1 3 5
some domain 1 1 5

• Focus curriculum guidelines for a variety of
  students
• Generated version 0 of a traceability matrix
• Numbers based on Blooms taxonomy

6
WG Quality

• Similarities/differences among quality attributes
• 3 essential elements stakeholder, system,
  concern
• QAs are all interrelated and there does not yet
  exist an orthogonal basis for defining QAs.
• Each QA has a cost and a value in achieving.
• QAs change form as you change abstraction level
• power (system level) ? idle mode (software side)
• Study measuring stability of architecture and
  relating it to causes and impacts would be
  useful. Stability could be caused by
• System death
• Getting architecture correct
• Fragility and fear of change

7
WG Architecting

• Agile methods pose neither new problems nor new
  solutions for architecting.
• Documentation
• Reverse engineering architecture from code will
  never suffice. Things such as rationale are
  simply not present in code.
• Cant rely on the architects memory, and video
  tapes, digital snapshots, etc. are not
  descriptive enough.
• There is no substitute for architecture
  documentation.
• Architects role The architect must code must
  be the first among equals in order to achieve
  good team morale.
• Patterns/styles help, need better tool support

8
WG Evaluation and Analysis

• Objectives Understand how to improve our ability
  to model and evaluate architectures.
• Representative issues
• How much architecture is enough? How many
  architects and architectural roles are needed in
  an organization?
• How do you identify the 5 top architectural
  issues for a system?
• How do we extract architecture patterns from
  code?
• How do we model architectures for
  mobile/pervasive systems?
• What properties do you evaluate? Why?

9
WG ADLs in Practice

• Why arent ADLs used in practice?
• Restrictive assumptions
• Single view(point)- oriented
• Lack of tool support
• Lack of domain focus
• Lack of rollout plan and tech transfer

• What to change?
• Linkage to technology
• Specialise with domain-specifics
• User studies
• what do architects really do/need?
• Simple tool support
• Visio, MagicDraw, Visual Studio, eclipse
• Graphical presentation support
• Improve tech transfer by aligning with trends

10
WG Documentation in Practice

• Enforcement/traceability Structuring/supporting
  architecture documentation for consistency of
  architecture implementation?
• Tooling What would a superb architecture
  documentation tool look like and what services
  would it provide?
• Multi-view consistency checking round-trip
  engineering to code
• Inhibitors why isn't documentation universally
  produced? What are the inhibiting factors?
• Need to focus on just in time architecture
  documentation
• What do I produce now, and for whom, to deliver
  best value?
• Documentation uses Can it ever take the place
  of a person?
• Does the role of documentation include protecting
  the project from architects who walk in front of
  busses?
• UML and architecture How could UML be a true
  ADL?
• Views as first-class concepts.
• A long list of other things. Should we worry
  about other languages on their own merits and
  translate those to UML?

11
WG Components and services

• Paper presentations
• Demonstration of (automatic) creation of
  connectors for different reasons (coordination
  protocols, providing different services, etc.)
• Which unwanted properties these connectors bring
  (performance, resource usage, timing behavior)??
• How to predict the system behavior when including
  such connectors?
• General discussion
• What characterizes differences between
  component-based and service-oriented approach?
  SOA objects, or not?
• How do we extend the principles (functional and
  non-functional aspects) of components to embedded
  systems?
• How do we get a software architecture that
  guarantees certain properties of a dynamic
  behavior?

12
WG Architecture Design Decisions

• High quality presentations discussions on
• tools processes for capturing architectural
  knowledge,
• prototyping processes
• case studies SMS to Air Traffic Control
• Key takeaways include need for
• common metamodel and language for expressing
  architecture design decisions (tools, reuse
  support)
• better understanding of casual relationships
  among decision decisions decisions interaction
  with environment.
• taxonomy of architecture decision decisions,
  using it as an aide for classification, ordering
  and socialization.
• Additional techniques and practices related to
  socialization, mentoring and communication
• Examples of good architectures are needed where
  architecture decisions are explicitly
  highlighted. Supports Boochs notion of a
  reading club

13
Technical blueprint vs leadership

• Dissonance SW Arch as technical system blueprint
  vs senior technical leadership
• Is the technical focus just out of touch, or
• Is the definition of SW Arch being diluted, or
• Is SW Arch in practice embedded in other
  responsibilities?

14
How can we resolve this dissonance?
Were looking at the bear and running faster
than the other guy
This is absolutely important and researchers
will help as much as possible, but we need to
bring in other researchers with skills in the
larger arena
15
Are we closing a Golden Age?

• Paul Clements invoked a golden age of aviation
• Golden Age of the Celestial Atlas rare book
  exhibit ?

• Dictionary.com any period (sometimes imaginary)
  of great peace and prosperity and happiness

16
Where Next?

• Move beyond the Golden Age
• It represents passage to maturity, not the end of
  research opportunities
• This groundwork shows where to make progress
• Each workshop topic
• Golden Ages are followed by development of
  richer, more broadly useful results
• Biplane ? 747, Airbus, spacecraft
• Illuminated maps ? Hubble photographs
• We need both for Mars Rovers

17
So, what are the SA opportunities?
18
So, what are the SA opportunities?

• Education
• Hans van Vliet and Mary Shaw
• Quality
• Len Bass
• Architecting
• Christine Hofmeister
• Evaluation Analysis
• David Garlan

• ADLs in Practice
• Eóin Woods
• Documentation in Practice
• Paul Clements
• Components and Services
• Ivica Crnkovic
• Architectural Design Decisions
• Jeff Tyree