Identifying refactoring through formal model based on data flow graph

Abstract

Identifying refactoring is an important step of bad smell removal process. Many developers have problem identifying refactoring, e.g., novice developers are not sure which refactorings should be applied. Even skilled developers may need to spend a lot of time doing this manually. Identifying refactoring techniques can alleviate these problems for developers. In our previous work, we proposed refactoring filtering conditions (RFC) that identify candidate refactorings for removing long method bad smell. These conditions analyze internal structure inside a method with data flow analysis. Our research improves these RFC based on more refined data flow analysis and presents its formal description. This research also demonstrates an approach of using the improved RFC of three refactorings. Moreover, we conducted an experiment to compare the efficiency of the results of our approach with that of our previous work and that of developer suggestion.