engleski

Processing and data collection of program structures in open source repositories

Software structure analysis with help of network analysis showed promising results in software engineering community. Some previous studies have presented potential of representing software structure as subgraph frequencies, that are present in software structure using network graphs, for effective program characterization and differentiation. One of the prerequisites for exploration of this potential is collecting large dataset with number of different software projects. Nowadays, there are plenty of open source code repositories which not only they contain source code, but also provide a lot of crucial information useful in guiding software engineering actions. However, systematical building of such large dataset is highly challenging and time consuming process. Therefore, our approach is to automate data collection process for open source repositories aiming to provide as large as possible dataset that could provide us reliable conclusions. In this paper we present software structure tool analyzer based on subgraph frequencies, discuss its automation opportunities and illustrate preliminary results obtained by its usage. Some potential research directions are discussed.

open-source repositories; automatic tool; software analysis

nije evidentirano