A Longitudinal, Multi-Project Study of Bug Tracking Productivity and Learning in Open Source Software Development

The dynamics of Open Source Software development have recently received a lot of attention from an organizational learning perspective. Following a network theoretic approach, we study the temporal development of communication network structure and productivity in order to detect associations among these constructs. Thereby, we identify a research gap in that related work either focuses on too few projects or utilizes insufficient numbers of analyzed timeframes. Our study is both multi-project and longitudinal in order to detect holistic influencing factors of successful Open Source Software development. First, we find that learning effects are present since productivity increases over time. Next, we observe that growth in team size impedes productivity whereas the continuous concentration on central nodes coincides with increasing productivity. Against our expectation, we also find that increasing centralization does not yield decreasing network density, possibly due to Open Source Software developers deliberately avoiding the dependence on bottleneck nodes.