摘要: 1 We construct a multi-relational, weighted, bipartite social network from activity data gathered from the Sourceforge. net research portal and Concurrent Version System (CVS) repositories. Using a novel clustering algorithm, we extract the social positions for the global social network as well as perform temporal analysis to show how the distribution of social positions change over time in the Sourceforge. net community. The social network analysis concept behind social positions is that people in the same position are indistinguishable from each other, and those people can be interchanged without altering the structure of the social network. We discovered that the extracted social positions differed between the global and temporal analysis of the social network; each providing different conclusions about the nature of individuals, and contrasting the two analysis techniques provides additional insights not obtaining by either single analysis. We discuss what these social positions mean in relation to our understanding of the organizational structure of open source software development projects.
参考文章(0)