Evolution of Collaboration in Open Source Software Ecosystems

Abstract

Although in the past decennium much research has been conducted in the field of software ecosystems, still little is known about how such ecosystems evolve. In existing literature, software ecosystems are often analyzed from a single point of time, without regard for their dynamic history.

We study an aspect of the evolution of open source software ecosystems: collaboration. In this thesis, we present a method to identify and analyze collaboration networks in software ecosystems, based on historical data from the online source code hosting service GitHub. We apply this method to analyze the evolution of collaboration within the well-established open source ecosystem around the Ruby programming language.

Our method can be used to identify and visualize collaboration networks around a given actor or software component over the course of time. This is done using statistical analysis techniques, without requiring subjective input. Our study analyzes over 200 collaboration networks to identify typical life cycle shapes and gives insight in the evolution an open source software ecosystem. Awareness of such collaboration networks and how they evolve is beneficial for actors in software ecosystems and for external stakeholders.