Exploring collaboration networks in open-source projects

Abstract

Analysis of developer collaboration networks presents an opportunity for understanding and thus improving the software development process. Discovery of these networks, however, presents a challenge since the collaboration relationships are initially not known. In this work we apply an approach for discovering collaboration networks of open source developers from Version Control Systems (VCS). It computes similarities among developers based on common file changes, constructs the network of collaborating developers and applies filtering techniques to improve the readability of the visualized network. We use the approach in case studies of three different projects from open source (phpMyAdmin, Eclipse Data Tools Platform and Gnu Compiler Collection) to learn their organizational structure and patterns. Our results indicate that with little effort the approach is capable of revealing aspects of these projects that were previously not known or would require a lot of effort to discover manually via other means, such as reading project documentation and forums.