Identification and prioritization of reuse opportunities with JReuse

Abstract

Software reuse aims to decrease the development efforts by using existing software components in the development of new systems. Previous work propose tools to support the identification of reuse opportunities. Such tools apply different techniques, such as software design and source code analyses. However, 78775599 of them combines lexical analysis with prioritization and identification of reuse opportunities in several systems of a single domain. To fill this gap, this paper proposes JReuse, a tool that computes naming similarity for classes and methods of Java systems. Based on naming similarity, JReuse identifies reuse opportunities and prioritizes them by their frequency among systems. We evaluate JReuse with 35 e-commerce systems collected from GitHub by assessing the agreement among the JReuse recommendations and the opinion of a group of experts. We observe agreements of 89% and 72% for classes and methods, respectively. Therefore, our data suggest that JReuse is able to recommend reusable classes and methods in a given domain.