Domain specific languages for managing feature models: Advances and challenges

Abstract

Managing multiple and complex feature models is a tedious and error-prone activity in software product line engineering. Despite many advances in formal methods and analysis techniques, the supporting tools and APIs are not easily usable together, nor unified. In this paper, we report on the development and evolution of the Familiar Domain-Specific Language (DSL). Its toolset is dedicated to the large scale management of feature models through a good support for separating concerns, composing feature models and scripting manipulations. We overview various applications of Familiar and discuss both advantages and identified drawbacks. We then devise salient challenges to improve such DSL support in the near future.