A longitudinal analysis of the dependency concentration in smaller modules for open-source software products

Recent studies on single releases of Open-Source Software (OSS) products have reported a higher concentration of dependencies in smaller modules. In this masters research, we investigated the findings of the recent studies by taking a longitudinal perspective and for a greater number of large-scale OSS products. Specifically, we investigated whether the reported concentration inequality always exists during product life time, and if so, how it changes over successive software releases. The results clearly confirm the earlier findings showing that, in all releases of all products examined, smaller modules were proportionally more dependent compared to larger ones. Generally, this concentration inequality increases over successive releases of the studied products. Since dependencies have been consistently associated with defects and failures in the software engineering literature, this comprehensive study of a large number of OSS products suggests that software teams should concentrate inspection and testing activities more on smaller modules. Our results also indicate that such focused QA efforts should be gradually increased as a software product evolves and matures.