Key points are not available for this paper at this time.
To build successful complex software systems, developers must collaborate with each other to solve issues. To facilitate this collaboration specialized tools are being integrated into development environments. Although these tools facilitate collaboration, they do not foster it. The problem is that the tools require the developers to maintain a list of other developers with whom they may wish to communicate. In any given situation, it is the developer who must determine who within this list has expertise for the specific situation. Unless the team is small and static, maintaining the knowledge about who is expert in particular parts of the system is difficult. As many organizations are beginning to use agile development and distributed software practices, which result in teams with dynamic membership, maintaining this knowledge is impossible. This thesis investigates whether emergent team structure can be used to support collaboration amongst software developers. The membership of an emergent team is determined from analysis of software artifacts. We first show that emergent teams exist within a particular open-source software project, the Eclipse integrated development environment. We then present a tool called Emergent Expertise Locator (EEL) that uses emergent team information to propose experts to a developer within their development environment as the developer works. We validated this approach to support collaboration by applying our approach to historical data gathered from the Eclipse project, Firefox and Bugzilla and comparing the results to an existing heuristic for recommending experts that produces a list of experts based on the revision history of individual files. We found that EEL produces, on average, results with higher precision and higher recall than the existing heuristic.
Shawn Minto (Sat,) studied this question.