Automatic Knowledge Sharing Across Communities: A Case Study on Android Issue Tracker and Stack Overflow. SOSE 2015: 107-116

OSSEAN: Mining Crowd Wisdom in Open Source Communities. SOSE 2015: 367-371

Abstract—Pull-Request (PR) is the primary method for code contributions from thousands of developers in GitHub. To maintain the quality of software projects, PR review is an essential part of distributed software development. Assigning new PRs to appropriate reviewers will make the review process more effective which can reduce the time between the submission of a PR and the actual review of it. However, reviewer assignment is now organized manually in GitHub. To reduce this cost, we propose a reviewer recommender to predict highly relevant reviewers of incoming PRs. Combining information retrieval with social network analyzing, our approach takes full advantage of the textual semantic of PRs and the social relations of developers. We implement an online system to show how the reviewer recommender helps project managers to find potential reviewers from crowds. Our approach can reach a precision of 74% for top-1 recommendation, and a recall of 71% for top-10 recommendation. http://rrp.trustie.net/ Keywords—Pull-request, Reviewer Recommendation, Social Network Analysis, Distributed Software Development