Empirical Evaluation of Reusability Models

Many research efforts have been directed into ways to quantify the degree to which a software component can be reused in other systems. As such, a number of different reusability assessment models have been proposed in the literature, taking into account several metrics that can affect reusability and different approaches to measuring it. In this paper, we conduct a longitudinal reusability assessment by applying three reusability models to a number of different projects (libraries and frameworks) with the goal of studying the long-term evolution of reusability in open-source software. The exploratory part of the study consists of reproducing and applying the chosen models on three different-sized projects for several released versions of the software and studying the transformations of reusability over time and how these relate to certain changes in quality factors or size of the software. Results show a more intense variation of reusability in earlier versions and more stable values towards later versions in applications, and a clear influence of complexity, modularity and cohesion on reusability scores.