Extracting communication structure of a development organization from a software repository

Abstract

Researchers have found that communication cost is one of the major overheads affecting the overall cost of software development. Actually, there could be communication problems caused by cultural difference, language barrier, different time zone, and, etc. in geographically distributed software development. Thus, extracting potential communication structural information is very useful in understanding and optimizing the development organization. Analyzing the communication structure is also crucial in order to resolve cost issues. While this is already true for general development organizations, geographically distributed software development organizations are especially sensitive to these issues because they are forced to rely on costly mediums. Therefore, this paper suggests a way to extract the development organization from software repositories with respect to the temporal locality. The temporal locality is important for when the project is prolonged for a lengthy period of time. In order to evaluate these issues, we define two metrics which measure contribution of the individual developer and communication need between developers. We also provide a tool to extract a communication structure using these two metrics. The extracted communication structure is visualized to give insight into managers and developers. Finally, we provide statistical results of empirical research to prove the soundness of our approach. The result shows that our approach reflects the real-world relationship between developers well.