To improve code structure by identifying move method opportunities using frequent usage patterns in source-code

Abstract

A smelly code is generally an indication of the poor quality of the software and it increases the understandability and maintenance efforts at the software programmer’s end. One technique to improve the quality is refactoring. Therefore, in this paper, we have identified the Feature Envy code smell and applied the corresponding Move Method refactoring. The code smell is tackled using the Frequent Usage Patterns (FUP’s) present in the source-code of the software. The FUP’s are identified at the method level and theyrepresent the set of member variables that are used by it either directly or indirectly. The identified FUP data is further used to cluster different methods using a newly proposed Clustering algorithm. Moreover, the proposed approach is successfully tested and evaluated on three standard open-source object-oriented software. The obtained results after evaluation confirm the ability of our proposed approach in enhancing the quality of the underlying software system.