A Measure of the Modularisation of Sequential Software Versions Using Random Graph Theory

摘要

Software module clustering is the problem of automatically partitioning the structure of a software system using low-level dependencies in the source code to understand and improve the system's architecture. Munch, a clustering tool based on search-based software engineering techniques, was used to modularise a unique dataset of sequential source code software versions. This paper investigates whether the dataset used for the modularisation resembles a random graph by computing the probabilities of observing certain connectivity. Modularisation will not be possible with data that resembles random graphs. Thus, this paper demonstrates that our real world time-series data-set does not resemble a random graph except for small sections where there were large maintenance activities. Furthermore, the random graph metric can be used as a tool to indicate areas of interest in the dataset, without the need to run the modularisation.