The Life Cycle of Features in Highly-Configurable Software Systems Evolving in Space and Time
summary
An automated mining approach processes full Git histories of four preprocessor-based HCSSs (13–20 years, 37,500 commits) to map features defined by external macros and compute change metrics (lines added/removed, scattering degree, tangling degree, toggling, files affected). Results show 3–7% of commits revise features (up to 8% touch only feature code), many revisions substantially alter implementation, and over 20% of features in some systems reach tangling ≥3, arguing for tool support and offering the mined dataset and scripts for follow-up studies.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 0:55Agent System Evolution
- 2:01Future Life Cycle
- 2:37Mining Approach
- 4:34Steady Goal
- 5:51Change Characteristics Matrix
- 7:19Change Characteristics Matrix Example
- 10:10Research Questions
- 14:18Summary
related talks
41mSoftware environmentalism
01h 35mImpact on maintainability and refactoring for higher-level design features
22mShifting Left, Shifting Right
46mMeet the Social Side of Your Architecture
49mTreat Your Code as a Crime Scene
42mFeature- and time-based software analysis - Balancing features, effort & quality
54mResolving technical debts in software architecture
30mAn Extensible Feature-Oriented Approach for Fine-Grained Code Quality Analysis