Reuse-Based Test Recommendation in Software Engineering

Still today, the development of e ective and high-quality so ware tests is an expensive and very labor intensive process. It demands a high amount of problem awareness, domain knowledge and concentration from human so ware testers. Therefore, any technology that can help reduce the manual e ort involved in the so ware testing process – while ensuring at least the same level of quality – has the potential to significantly reduce so ware development and maintenance costs. In this dissertation, we present a new idea for achieving this by reusing the knowledge bound up in existing tests. Over the last two decades, so ware reuse and code recommendation has received a wide variety of a ention in academia and industry, but the research conducted in this area to date has focused on the reuse of application code rather than on the reuse of tests. By switching this focus, this thesis paves the way for the automated extraction of test data and knowledge from previous so ware projects. In particular, it presents a recommendation approach for so ware tests that leverages lessons learned from traditional so ware reuse to make test case reuse suggestions to so ware engineers while they are working. In contrast to most existing testingassistance tools, which provide ex post assistance to test developers in the form of coverage assessments and test quality evaluations, our approach o ers an automated, proactive, non-intrusive test recommendation system for e icient so ware test development.