Towards Knowledgebased Product Derivations

1
Towards Knowledge-based Product Derivations
Groningen, The Netherlands
3.12.2004
Dr. Martin Becker,Michalis Anastasopoulos, Franz
Grzeschniok
Sauerwiesen 6 D-67661 Kaiserslautern Germany
2
Outline

• Motivation
• Approach
• Product Derivation
• Required Knowledge
• Derivation Process
• Conclusion

1/12
3
Motivation

• Overall ROI of a Product Line heavily depends on
• reuse achieved in application engineering
• Reuse is not an option it has to happen
• Reuse of existing solutions has to be efficient
  and effective
• Impact of reuse has to be considered (e.g. to
  provide feedback)

• Current situation
• Increasing interest in Software Product Line
  approaches
• Considerable BOK exists for domain engineering
• Situation for application engineering is
  different
• many implementation level approaches with focus
  on technical details
• lack of standard vocabulary
• difference between software development and
  product derivation is unclear
• Experienced problems
• Organizations are missing well-defined
  application engineering processes
• Missing knowledge on how to derive products from
  a reuse infrastructure in an efficient and
  effective way
• Institutionalization and automation of these
  processes is hard

2/12
4
Approach

• Our Goal Enhance the efficiency of product
  derivation
• Contribute to a sound methodological basis for AE
  
• Foster the automation of product derivation
• Combine the knowledge-based approaches developed
  in ConIPF with SOTP Tools
• Followed Approach
• Identifi the necessary knowledge for product
  derivation in existing PLs
• Define models to capture this knowledge and
  illustrate their application
• Analysis of SOTP tools how this knowledge can be
  processed
• Extension and integration of SOTP tools to
  support a (semi-)automated product derivation
• Focus of presentation/paper is on step 1 2

3/12
5
Product Derivation vs. Software Development

• Product Derivation ! (Software Development
  Reuse)
• Impact of systematic reuse is more substantial
• Impact of reuse depends on
• structure of reuse infrastructure
• development strategy top-down or bottom-up
• provided reuse support
• deployment of reuse support

4/12
6
Required Knowledge
5/12
7
Knowledge Types

• 1. ConceptsProperties, Specifications
• 3. Intermediate Concepts lt-gt Relation
• 2. RealizationsArtefact, Artefact Elements

6/12
8
Overall Knowledge Model
Feature Models can describe Artefacts
Hierarchical structure of Assetelements/Assets
7/12
9
Intermediate Knowledge
How and where to capture in SOTP toos?
8/12
10
Knowledge Models
ConceptualKnowledge
Implicit
OverallKnowledge
Product LineModel
Explicit
FeatureModel
Intermediate
RefinedProduct LineModel
9/12
11
Derivation Process
Recurring process pattern
10/12
12
Conclusion

• PLE strongly affects product development
• Lack of defined AE processes and knowledge
  causes substantial inefficiencies ? reduced ROI
• In this paper we have
• identified the required knowledge and and its
  application
• proposed a set of models to capture the knowledge
• presented a common process pattern
• Further work
• Support AE with SOTP tools
• What you can do?
• Read the paper
• Give us feedback

11/12
13
Thank You
12/12