Detecting and Recommending Mitigations for Impactful Risky Software Changes

点击翻译按钮获取中文摘要

Software systems play an important role in our daily lives, making the quality of software systems of paramount importance. To ensure the quality of software systems, a large amount of research focused on the detection of source code packages, files and changes that contain defects. However, the adoption and usefulness of such approaches in practice remains limited. Some of the most commonly cited reasons for the lack of adoption are 1) metrics used in the detection are not actionable, 2) the impact of defects is not taken into consideration and 3) no guidance is given on how to mitigate the detected defects. The aim of the proposed research is to address the aforementioned limitations, making defect detection techniques more pragmatic. To do so, we plan to propose models that use actionable metrics to detect impactful risky changes. Furthermore, we will use historical data and employ qualitative and quantitative approaches on the fixes of the impactful risky changes in order to propose mitigations for the detected impactful risky changes. The successful completion of this proposed project will enable software researchers and practitioners to better understand the key development factors that enable pragmatic defect detection. The novel contributions of the proposed research are: 1) the proposal of models that use actionable metrics to effectively detect impactful risky changes and 2) a catalog of mitigations that can be applied to address these impactful risky changes. Large-scale empirical studies will be performed using large open source and industrial projects, to determine the effectiveness of the proposed models, actionable metrics and proposed mitigations. The proposed research will have a direct impact on the software engineering research and practice since it will advance the field by providing pragmatic detection techniques, proposing novel mitigation strategies and providing an extensible tooling framework. Furthermore, the proposed research will expose, train and enable highly qualified personnel (HQP) to contribute to the state-of-the-art in software quality research.