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 insucient numbers of analyzed timeframes.
Our study is both multi-project and longitudinal in
order to detect holistic in
uencing factors of successful Open
Source Software development. First, we nd that learning
e ects 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 nd that increasing centralization does
not yield decreasing network density, possibly due to Open
Source Software developers deliberately avoiding the dependence
on bottleneck nodes.