Discovering Community Types in Open-Source: A Systematic Approach and Its Evaluation

The open-source phenomenon has reached considerable proportions to the point in which it is virtually impossible to nd large applications that do not rely on open-source components. However, such grand adoption may turn into a risk if the community and organisational aspects driving open-source work (e.g., contribution policy or release schemes) are left implicit and untracked. We advocate the explicit study and automated support of such aspects and propose Yoshi (Yielding Open-Source Health Information), a tool able to map open-source communities onto known organisational and social structure types (e.g., Formal Communities characterized by a very formal community organisation). This mapping is bene cial since it allows (a) further investigation of community health using metrics from organisations research, (b) reuse of type-speci c best-practices from the same literature, and (c) diagnosis of organisational anti-patterns. We evaluated the tool in an empirical study involving 25 open-source communities hosted on GitHub. We evaluate (i) whether the metrics computed by Yoshi represent valid measures to characterise software communities, (ii) the extent to which the tool is able to correctly estimate the structure of a community, (iii) whether community smells (i.e., socio-technical anti-patterns) can be diagnosed using Yoshi’s typi cation, and (iv) whether such typi cation can help in understanding the health status of the repository. We conclude that the tool o ers a valuable basis to monitor key community traits behind open-source. We made the proposed tool publicly available and open-source.