This project will investigate how, when and why open source collaborative software development projects succeed or fail. Open source software runs the Internet, dominates the web server market and competes successfully in operating systems and even applications markets. This success is remarkable given the major communication and coordination problems that typically plague other forms of geographically distributed collaboration. But despite the visible successes of open source, it is also true that most projects fail. This project will conduct theory-driven empirical studies of coordination, social networks, and outcomes. The first step will be to replicate studies of commercial software development projects to determine if coordination in open source follows the same processes, or if it is fundamentally different. The next step will be to perform comparisons of open source and commercial developments to see if the theoretically-predicted differences are observed. Next, a computational model of open source software development will be constructed. Model predictions will be validated against actual observations. Finally, field experiments will be conducted to make specific coordination improvements in selected open source projects.
