Automatic Identification of Cohesive Structures within Modularity Reengineering

Abstract—The quality of software systems depends heavily on the quality of their structure, which affects maintainability and readability. To improve the quality of structure, a system can be restructured. This paper describes a restructuring pro-cess, which uses a combination of strongly connected compo-nent analysis, dominance analysis, and intra-modular similarity clustering to identify and preserve structures that have been thoughtfully placed together, but would be separated by pure metric-based or similarity-based techniques. The use of the proposed method allows a significant reduction of the number of components that should be moved. Therefore, the number of false movements is alleviated. The proposed approach was implemented in a prototype and illustrated by statistics and examples from 18 open source Java projects. A coherence metric is introduced to further improve restructuring results. Keywords-Source code organization, Restructuring, reverse engineering, and reengineering, Metrics I