Learning from Contributions
Social coding is a software development approach that provides a collaborative environment to developers, encouraging them to share and discuss new ideas and knowledge. Such collaborative environments implement a communication channel between developers and software users, enabling them to contribute to products by acquiring what their users really think of their products, which aspect they care about the most, what functionality they are still expecting, etc. There are many shapes of contributions to software development (i.e., pull-request, mailing lists, commit logs, issue reports, RFCs, social media, etc.). These contributions generate a high volume of communication data in the software development tasks such as documentation, internationalization, reporting an issue, adding some tests, reproducing a bug, fixing a reported bug, adding requested feature, removing unused or redundant code, etc… Such communication data exists mainly as unstructured data. Extraction and processing of unstructured data are challenging.
In this project, our objective is to gather and analyze various communication data in software development towards learning contributions and relate them to the characteristics of the software products and processes.