Published in Lecture Notes in Computer Science vol. 10826: “Supporting Product Line Adoption by Combining Syntactic and Textual Feature Extraction” (Kicsi, Vidács, Csuvik, Horváth, Beszédes and Kocsis).
The work deals with moving an existing legacy system to a software product line architecture while keeping the variants in one code base. It proposes an automatic method that links high-level features defined by domain experts to program code, combining call graph information with the textual similarity between code and features. The setting was an industrial product line adoption project for a large logistics information system written in the Magic 4GL language.
Read the paper: publisher page
