Transcript Evolution Group 2

Group Members
• Infrastructure/Technology Subgroup
•
•
•
•
•
Kentaro Yoshimura (Hitachi, Japan)
Uirá Kulesza (UFRN, Brazil)
Sergio Soares (UFPE, Brazil)
Cheng Thao (UW-Milwaukee, USA)
Ethan Munson (UW-Milwaukee, USA)
Evolution: Infrastructure
Main Objectives
• Assessment testbed and tools
• Common data for many studies
• Shared results from studies for comparison
• Evolution histories
• In order to learn about practices and needs
• Identify evolution patterns
• Refactoring support for SPL
• Splitting/merging features
• Support for large scale product lines
• Many, many features
Evolution: Infrastructure
Main Objectives (2)
• Real CM support for variation
• Is this necessary?
• Perhaps not when feature/decision models are used
correctly
• What does empirical evidence say?
Evolution: Infrastructure
Matches
• Industrial experience
• Yoshimura/Li/Krueger with everyone
• Evidence that can be mined to motivate
refactorings or version control approaches
• Many practical issues that arise from large scale of
feature set and product line
• Empirical and case study support
• Soares and Leuven group provide data
• Used by Munson/Thao and Kulesza
• Yoshimura/Li/Krueger can help us make sure that
testbed is realistic
Evolution: Infrastructure
Gaps
• Integration testing support
• Lack of hierarchical structure to feature
combinations means that no one unit is in
charge of each integration
• Legacy systems
• Mining features from existing code
• Conditional compilation  features?
• build scripts  products & components
Evolution: Infrastructure
Possible Future Results
• Assessment testbed and tools
• Common data for many studies
• Shared results from studies for comparison
• Product line operations or refactorings
• Configuration management support for
SPL
• Direct CMS support for impact analysis
Evolution: Infrastructure