Over the last decade, significant theoretical and empirical research, led by scholars such as Yochai Benkler (The Wealth of Networks), Eric von Hippel (Democratizing Innovation), and Steven Weber (The Success of Open Source) has been conducted about collaboration in open source software, and more generally, collaboration in open content. This paper builds upon these and other important work and undertakes original empirical analysis of data from Sourceforge.net, to begin to identify "design principles" for free/libre and open source software (FOSS) collaborations. This paper will summarize theoretical and empirical results to date of a research project on FOSS collaboration sponsored by the U.S. National Science Foundation. The broad goal of the research project is to identify factors that contribute to the success or abandonment of FOSS projects.
The first section of the paper will conceptualize FOSS projects as a "commons," and will provide a more complete picture of what we refer to as the "Open Source Ecosystem." What is becoming clear is that FOSS is no longer always an "all-volunteer" effort. The situation is becoming more complex, with inclusion of industry, government organizational interests, and the entrance of nonprofits and nonprofit foundations.
The second section of the paper will provide a framework for analyzing open source software commons, and will describe factors, grounded by theory and empirical research, that are thought to influence the success or abandonment of these projects. This section builds upon over forty years of research studying commons situations, many of which fall in natural resource settings. We will build on this and present a summary of factors which, theoretically, empirically, or both, are thought to influence the success or abandonment of FOSS projects.
The third section will turn to an empirical analysis of FOSS projects, based primarily on analysis of FOSS project metadata acquired from the open source hosting site Sourceforge.net (SF). We see SF, and metadata repositories that are storing longitudinal historical data of SF such as FLOSSMole at Syracuse University, and the SF data repository at the University of Notre Dame, as "a remote sensor" of FOSS. Given that this hosting site is likely to be around for years to come, it is important that we understand what we can learn about these projects from the metadata it maintains.
The fourth section will provide some reflections on the above analysis. We will include discussion not only on the implications for open source software projects, but also more generally for open content collaborations. It is here that a direct connection to the FOSS4G conference will be made. The primary author is the current chair of the OSGeo educational committee, and is working to initiate an open content collaboration in FOSS GIS educational material.
While this paper does not focus specifically on open source Geographic Information Systems, we expect the theoretical and empirical work will be of particular interest to many in the FOSS4G community, and, as we just mentioned, has relevance to OSGeo's educational initiatives. Finally, having an opportunity to present the work to date to the FOSS4G community - people directly involved in open source collaboration -- will provide valuable feedback for our continued research.