Freedom and forking in open source software: the MariaDB story

While significant factors that affect the open source community’s interest to participate in a development project have been studied, there has been little focus on the motivating factors that can cause a contributor to become a competitor by utilizing the right to fork a program i.e., to copy an existing program’s code base and use it to begin a separate development.

The right to copy an existing program’s code base and use it to begin a separate development is guaranteed by all open source licenses. However, this right to fork a program is rarely exercised. Indeed, there is strong social pressure against code forking stemming from the negative side effects of code forking, such as conflict and duplicated efforts among developers.

This paper details the events that led Widenius, the founder of the MySQL project, to decide to fork MariaDB from MySQL. Our findings confirm the previously held notion that there is a high threshold for starting a competing fork. While the few studies that exist of competitive forks find the reasons to be due to disagreement among developers, in the case of MariaDB the fork was caused by Widenius’ concerns regarding the uncertainty of the future freedom and openness of the MySQL codebase.

This article makes three contributions. Firstly, it further validates the existing notion that there is a strong threshold to starting a competing fork. Secondly, it offers an in-depth analysis of the events and motivations behind the birth of a fork. Thirdly, it contributes to theory by introducing the freedom factor hypothesis: limiting either developers’ freedoms to contribute to a project or the freedom inherent in a project's license increases the likelihood of a fork.