Software Superfamilies Based on Sub-Graph Significance Profile

The significance of triad appeared in open source software is studied. It is found that the local structure trends to be networked with the increase of software scale. Tree style sub-graph trends to decrease, but most of close style sub-graph trends to increase. After comparing triad significance profiles and correlation between open source software networks, we have found that software networks could be divided into 3 clusters which are consistent with 3 of 4 well known super-families contain different networks from various domains. Most software networks have similar local structure with biological networks. It seems that software scale may be one of the reasons causing different sub-graph significance profiles. With the increase of software scale, the significance profile of triad trends to be similar. The experiment results show that the software network topology is very similar to biological networks, part of the research methods, research results and evolution mechanism of biological network are very helpful to the research in software network. The experimental results also show analyzing software networks, in particular to analyze the small-scale software network, that should be the first to analyze the local structural features. Depending on the local characteristics, it is necessary to take a different research approach and research strategy, not simply be treat them equally.