Automatically Generated Feature Model from Requirements: Toward an Enhanced Formalism

The feature model is one of the most important concepts in software product line development process, as it helps to represent commonalities and variabilities among software products. Several research works are currently conducted to generate a feature model of a software product line from its requirements document. However, the generated feature models do not include data and actors aspects which are essential parts of any software product line. The work presented in this paper is part of an automated feature model generating project which explores suitable input requirements document notations, simplifies the generating process, and outputs generated feature model formalism. It aims to propose a powerful feature model that should be the target of any automated process generating a feature model. Thus, in this work, different recent techniques that automatically generate feature models are analyzed. The lack of data and actors in these generated feature models was stated, and a feature model, aimed to be a good generated target, integrating data and actors with other conventional features is proposed. A case study evaluating the proposed target feature model formalism is presented.