Skip to main content

Social Network Analysis of OSS

Mozilla Firefox, Apache, Linux, Android, W3C, MySql, etc. all these software solutions that we use widely today have one thing in common: they are all open source projects. Today, there are close to 5000 active and popular open source software projects which are revolutionizing the software world in a big way. It has been recorded that the use of such software solutions has resulted in consumer savings of about $60 billion per year. Thus, OSS has become an integral part of the infrastructure of modern society and this makes it interesting to understand how it developed. OSS development community is classic setting to understand collaborative social networks. Collaborative social networks are made up of individuals and organizations that are linked to each other and form one large cluster with some particular characteristics.  The network model in this scenario has nodes that are organizationally and geographically distributed and is purely online based. In particular, there are questions raised about the motivations of users to voluntarily participate, their similarity to other online general network structures like that of Face book and their evolution over time.

This OSS community when compared to other community networks has some interesting insights. A number of studies have been conducted with the data collected from SourceForge. It was shown that this network structure follows a power law distribution. Power law distribution is having a few nodes with very high degree and a large number of nodes with low degree (number of edges). This can be considered relevant since these communities start in a self organizing process and nodes have preferential attachments. They have the freedom to choose their projects and enter and exit at any time. A java developer will be most likely linked to another java developer and not a Perl developer. The network structure in developer network can be understood in two ways: one where nodes in the network are developers and other where nodes are projects. The network structure in both these conditions follows the power law distribution.

It is also observed that the nodes in such networks have smaller degrees of separation and are more closely linked to each other than in a general social network. They follow the small world phenomenon. This could be attributed to the fact that they are smaller range of topics in such a community than in a social networking community. Modularity or the strength of the community structure is high in these networks suggesting that a close knit community and they steadily increase with time. Studies have shown that the existence of small world phenomenon in this structure leads to the rich-get-richer effect where some projects are more attractive to developers than others and thus leads to their success.

Knowledge of such network structure can help in effective team building and coordinating processes in organizations which are moving towards virtual teams. Study of such network structures are a good predictor of group behavior, trust, knowledge generation, and information diffusion. There are still many unanswered questions in this field like the explanation of dynamic life cycle, sustainability of such a structure, validity, etc.



Leave a Reply

You must be logged in to post a comment.