Utilizing a multi-developer network-based developer recommendation algorithm to fix bugs effectively
Geunseok Yang, Tao Zhang, Byungjeong Lee · 2014
Recently, bug fixing has become an important part of software maintenance. In large-scale projects, developers rely on bug reports to guide any bug-fixing activities. Due to a great number of bug reports submitted into the bug repository, the workload of the triagers who are responsible for arranging developers to fix the given bugs is very high. In order to reduce the triagers' workload, a number of approaches (e.g., machine learning algorithms and social network metrics) were proposed to study who should fix the bug report. In this study, we propose a novel algorithm for developer recommendation. We first introduce a component and a similar bug-based selection process to verify the candidate fixers, then by adopting the number of comments and commits, we construct a multi-developer network so that ranking these candidates for finding the most appropriate fixer to resolve the given bug. In order to evaluate our work, we measured the effectiveness of our approach based on 3,008 bug reports from the JBoss Issue bug repository. We also compared the proposed approach to three previous studies. The result shows that our approach performs the task of bug triage effectively.