MAVIS - MAnagement of Variability in Information Systems

点击翻译按钮获取中文摘要

The evolution of software systems becomes increasingly complex when they are customized for diversemarkets and application contexts. Much of this arising complexity stems from unmanaged variability of systemfunctionality and features. Unmanaged variability impedes testability and quality assurance and increases thelikelihood of defects. The software engineering community has developed well-founded practices to developsoftware-intensive systems with variability. Such so-called software product families (or software product lines- SPL) are evolved and quality assured based on architectural models and methods that carefully distinguishbetween core functionality and variant features."Legacy" systems that have not been engineered with SPL practices must be reengineered as a prerequisite toimplementing SPL practices. This involves the (1) extraction of the software's variation points, (2) theassociation of software artifacts to core components and variants, and (3) the refactoring of the legacy systemto facilitate SPL practices.The research community has developed methods and tools in support of migrating legacy software to SPLs.However, the applicability of these results is often limited to particular kinds of systems. Few empirical studiesexist to assess the maturity and readiness of SPL reengineering methods and tools "at scale" and withindustrial-strength case studies that were not specifically chosen to demonstrate a given method.The objective of this project is to review, select and assess the readiness of current SPL reengineeringmethods and tools in the context of a heavily evolved software system in the medical domain, developed by ourpartner OSCAR EMR. The results of this project will provide novel empirical insights on the applicability,maturity, scalability and limitations of current SPL reengineering methods. They will benefit OSCAR EMR aswell as the research community, as they will provide an empirical basis for technology transfer and follow-upcollaborative research on improving existing methods and tools.