Boundary Spanners in Open Source Software Development: A Study of Python Email Archives

In many open source software development communities, a significant proportion of development is undertaken by a relatively small number of individuals, the "core members". The stability and longevity of this group of most active developers are crucial for the success of the project. While there has been prior work on identifying key individuals in open source development, little attention has been devoted to the identification of cross-cutting core individuals (boundary spanners) whose responsibilities span across different functional areas of open source development (e.g., who are involved both in development-centric activities and user-centric activities). To address this gap, we propose an approach to identify the core cross-cutting members and their roles within the community through analyzing email communication repositories. We use Social Network Analysis (SNA) tools to identify the most active core members in different forums (that have different focus such as Python-dev that focuses on language evolution and Python Lists that focus on user support), and their activities over time, thus identifying the core developers and their involvement in different community mailing lists. Based on the involvement of a core developer and the overall social structure of the network of core developers, we also present an approach for identifying a potential replacement for a community administrator that steps down. Using email repositories of six main Python forums as the case study domain, we computed several social network analysis metrics to characterize the core developers and their importance in the Python community.