Migrating the Android Apo-Games into an Annotation-Based Software Product Line

Abstract

Most organizations start to reuse software by cloning complete systems and adapting them to new customer requirements. However, with an increasing number of cloned systems, the problems of this approach become severe, due to synchronization efforts. In such cases, organizations often decide to extract a software product line, which promises to reduce development and maintenance costs. While this scenario is common in practice, the research community is still missing knowledge about best practices and needs datasets to evaluate supportive techniques. In this paper, we report our experiences with extracting a preprocessor-based software product line from five cloned Android games of the Apo-Games challenge. Besides the process we employed, we also discuss lessons learned and contribute corresponding artifacts, namely a feature model and source code. The insights into the processes help researchers and practitioners to improve their understanding of extractive software-product-line adoption. Our artifacts can serve as a valuable dataset for evaluations and can be extended in the future to support researchers as a real-world baseline.