دانلود رایگان مقاله مدیریت دانش در جوامع OSS: رابطه بین ساختارهای شبکه متراکم و پراکنده

ABSTRACT

Some authors in the literature have addressed knowledge transfer via weak ties between organization’s units which are themselves strongly tied inside (e.g. Hansen, 1999). Some others have investigated knowledge management among open-source-software (OSS) developers and discussed factors influencing knowledge transfer within development teams (e.g. Joshi and Sarker, 2006). In the domain of open source software (OSS) communities, more companies are now attempting to establish relationships to benefit from these potential value-creating communities; and project managers could in fact target different goals within software development teams including knowledge transfer within and between teams. We step forward to distinguish knowledge transfer within groups as opposed to knowledge transfer between groups; where relevant projects are bundled into separate strongly intra-connected groups. In knowledge management literature there is a trade-off between sparse network structures (Burt, 2000, 2002) versus dense network structures (Walker et al., 1997; Coleman, 1988). It is argued that the former facilitates the diffusion and generation of ideas among groups, while the latter affects the implementation of idea within each dense group. To our best knowledge, there has been no study to investigate the relationship between dense and sparse network structures. We propose that knowledge transfer within dense groups has a positive influence on knowledge transfer between sparse groups, in that intragroup density, group size, developers centrality and betweenness could impact intergroup coupling. To prove our hypothesis, we use a complex network of open source software (OSS) as the domain of interest, where developers represent nodes and two developers contributing to a project task represent a network tie. Developers contributing to tasks in groups other than their own can explore novel ideas via sharing knowledge, whereas developers contributing to tasks inside groups exploit ideas to improve those projects. We investigate the idea both analytically and empirically within 4 months, 8 months and 1 year lagged time, and finally show that intragroup density has a positive whereas developers’ centrality has a negative influence on intergroup coupling.

4. Conclusion

Knowledge management among open-source-software (OSS) teams has been noticed in the literature e.g. Joshi and Sarker (2006) discussed factors influencing knowledge transfer within development teams. Ojha (2005) also discussed knowledge sharing between team members based on similarity-attraction paradigm. In this study, we attempted to distinguish between the factors affecting knowledge transfer within groups as opposed to between groups. To our best knowledge, there has been no study to investigate the relationship between dense and sparse network structures and whether knowledge transfer in dense network (inside groups) has an influence on knowledge transfer in sparse network (between groups). This study demonstrates that knowledge transfer within groups could influence knowledge transfer between groups. In order to investigate how intragroup density affects intergroup coupling, we used the link cost/benefit method introduced by Jackson and Wolinsky (1996) using utility function for each project based on benefit and cost of new link formation. We showed that when initial link is formed between two groups, subsequent link formation is always cost-wise beneficial to be formed, when there is a higher intragroup density. One can explore this more, and investigate the causal relationship in more detail in that how dense a group should be? Is there a threshold, upon that all intergroup connections are automatically formed; and below that threshold, no extra intergroup connection is formed. Jackson and Wolinsky (1996) discussed unique strongly effi- cient network where for the case of a complete or fully connected graph, ρ − ρ2 > c should hold true, however for a condition that all nodes of a network is directly connected (a start encompassing everyone) but is not a complete graph, then ρ − ρ2 < c< ρ + ((N-2)/ 2)ρ2 should hold true. This is the threshold which shows how dense a group should be, within which all nodes are directly connected.