MS Thesis opportunity!
Feature model extraction from subversion repository
Description:
Context: Given the high variance in the manufacturing industry, software has to work with all the possible hardware configurations. Software product line (SPL) is one solution to achieve this. The implementation of SPL can vary from company to company but one of the most widely used implementations is the 150% model. The 150% model contains some extra functionality that is removed/configured to derive new products.
Motivation: All the artifacts of the 150% SPL ends up in the code repository. Most companies find it hard to co-evolve the feature model and a feature model describing the decisions for a specific variant. Automated approaches for co-evolution of feature models are required.
Method and Results: A tool supported method should be developed that work with SVN repository to extract the feature model. The approach method should heavily use directory structure (for analyzing what features were removed or added), dependencies, commit data (to extract when a feature was introduced and by whom), and, Simulink models’ metadata (to extract human readable feature names and how features were configured) to extract the feature model. The method is supposed to be a continues-co-evolution method deployed as Jenkins plugin.
Expected Outcome: 1) algorithm(s) and Jenkins plugin(s) to analyze and detect dependencies of different SVN branches (typically consist of different Simulink models), and 2) generating FeatureIDE based feature models.
This thesis is defined as part of the XIVT project (an ongoing research project on variant testing)
Contact Person: Muhammad Abbas (muhammad.abbas@ri.se) OR Mehrdad Saadatmand (mehrdad.saadatmand@ri.se)
RISE SICS Västerås