Duplication detection when evolving feature models of software product lines

After the derivation of specific applications from a software product line, the applications keep evolving with respect to new customer's requirements. In general, evolutions in most industrial projects are expressed using natural language, because it is the easiest and the most flexible way for customers to express their needs. However, the use of this means of communication has shown its limits in detecting defects, such as inconsistency and duplication, when evolving the existing models of the software product line. The aim of this paper is to transform the natural language specifications of new evolutions into a more formal representation using natural language processing. Then, an algorithm is proposed to automatically detect duplication between these specifications and the existing product line feature models. In order to instantiate the proposed solution, a tool is developed to automatize the two operations. © 2015 by the authors.