An Approach to Suggest Code Smell Order for Refactoring

Abstract

Code smell is an indicator of issues in source code qualities that may hinder maintenance, and evolution. Source code metrics are used to measure the quality of the code. In the literature, there are many code smells, refactoring techniques, and refactoring tools. However, a software project often contains thousands of code smells and many of them have no relation with design quality. It is a challenge for developers to decide which kind of code smell should be refactored first. We have proposed an approach that suggests a code smell order based on two aspects: (1) finding relevant metrics for each code smell dataset with the help of feature selection technique (2) analyzing the internal relation among the code smells with those relevant metrics. With this analysis, we are suggesting code smell order for developers to save their effort in the refactoring stage. The suggested order is evaluated on simple java source code.