The Governance of Open Source Software Development Projects
Di Tullio, Dany
MetadataShow full item record
This thesis investigates the following research questions: (1) What is open source project governance and how can it be conceptualized? (2) What is the relationship between the dimensions of OSS governance and the specific purposes that governance is hypothesized to serve in open source projects? (3) How do the major configurations of governance dimensions affect the performance of open source projects? Two studies were conducted to answer these questions: an exploratory qualitative study and a survey study. In the qualitative study, we clearly defined, developed, and validated the various dimensions of OSS governance. This allowed for the identification of a limited number of configurations of governance dimensions that most frequently occur in open source projects. We found that a patterning in governance dimensions takes place because dimensions are in fact interdependent. Therefore, only a fraction of the theoretically conceivable configurations of governance dimensions appear to be viable and were observed among a range of open source projects. This provided us with a preliminary understanding of how these dimensions configure to create three distinct configurations of project governance which were labeled as follows: Open Communities, Managed Communities, and Defined Communities. In the quantitative survey, we first validated these configurations of governance using cluster analysis and then tested the relationships between these configurations (clusters) and the specific purposes that governance is hypothesized to serve in open source projects, namely solve collective action dilemmas, solve coordination problems, and create a climate for project excellence, while assessing their influence on the performance of projects. The results confirmed the presence of three main governance configurations (clusters) and also showed that open source projects that adopt a Defined Community approach to governance were the most successful. In these types of projects, the combination of a tightly managed software development process with a decentralized community management structure was found to create a balance between anarchy and control that allows these projects to benefit from one of the virtues of open source development, the open contribution and participation of a wide variety of talented developers, while avoiding the pitfalls of an uncontrolled and scattered development process.