Automated analysis of dependent feature models
Reimar Schröter, Thomas Thüm, Norbert Siegmund, Gunter Saake · 2013
Feature models specify valid combinations of features in software product lines. With dependent feature models (DFMs), we apply separation of concerns to feature models for two main benefits. First, we can modularize feature models into parts relevant to groups of stakeholders. Second, we are able to model dependencies between different software product lines in a multi-product-line scenario. To ensure consistency and correctness of DFMs, we have to apply analyses, such as dead-feature detection. We discuss why DFMs challenge the detection of inconsistencies, present how to reuse existing analyses for DFMs, and propose new analyses to supplement existing ones. We apply automated analyses in five steps and evaluate the approach using DFMs specified in VELVET by our prototype VeAnalyzer.