This is my literature review of a paper by G. Hertel et al. titled – “Motivation of software developers in Open Source projects: an Internet-based survey of contributors to the Linux kernel”.
Based on discussions in the mailing list, the two motivations for contributing to Linux kernel were identified: “fun to program” and to compete with other projects and commercial software. But since there wasn’t any empirical evidence to verify this anecdotal evidence, the authors conducted a research study the Linux Kernel community to verify the claim.
Open Source Software (OSS) development has some properties of a social movement and a small work team. The authors used models developed by sociologists, to explain participation in social movements (EKM) and work teams (VIST). Thus the authors combined the data from three sources – Open Source community, research on social movement participation and work teams to develop a model for explaining OSS participation.
The authors announced the survey on the mailing list. In the spirit of openness the survey and its motivations were published on a website and invited comments. The final survey was expected to take a 15 minutes to complete. Survey respondents were classified into developers (kernel developers or module maintainers) and readers.
Using factor analysis, seven factors were identified in the responses as follows:
- Identification as Linux user
- Identification as a Linux developer or with a Linux subsystem
- Pragmatic motives related to the improvement of one’s own software and career advantages
- Norm-oriented motives related to reactions of relevant others (family, friends, colleagues)
- Social and political motives related to supporting independent software and networking within the Linux community
- Hedonistic motives such as pure enjoyment of programming
- Motivational obstacles related to time losses due to Linux-related activities
Not all the factors had influence on the contributions made to the project. Only the following factors predicted contributions to Linux Kernel:
- Factors (2) and (7) predicted the numbers of hours spent, especially by developers.
- Factor (3) predicted willingness to engage in Linux related activities in future
The developers worked on subsystem like a virtual team. Survey respondents who were developers answered additional questions about their contribution like number of lines and patches. These variables were predicted using motivational processes.
- Number of hours spent is correlated with how important a developer considered his/her contribution to the subsystem (referred as Instrumentality).
- A developer’s willingness to increase participation in the subsystem was correlated with his/her perceived value of the subsystem goals (referred as valence) and instrumentality.
- Number of patches accepted from a developers was explained by self-efficacy and instrumentality.
- Finally, the number of lines a developer submitted was correlated with self-efficacy.
The authors have made a good case for treating open source community participation as a social movement. Which gives a good source of reference for researchers studying open source communities. The study also gives a set of motivational factors that can be used to other open source communities.
I was concerned that survey might have selection bias since respondents choose to participate. Also the paper did not mention the population size, although that is difficult to determine given the nature of the medium. It is not surprising that the authors did not receive many responses when they started a discussion about the survey before launching. I’m curious if the researchers received many comments that were against the intentions of the study, how they would proceed. Finally, I was tad disappointed that the research did not use any data from mailing list or source code repository. It would have been interesting if the study involved correlating surv