PROJECT TITLE :

Mining Version Histories for Detecting Code Smells

ABSTRACT:

Code smells are symptoms of poor style and implementation choices that may hinder code comprehension, and presumably increase change- and fault-proneness. Whereas most of the detection techniques just rely on structural info, several code smells are intrinsically characterised by how code components modification over time. In this paper, we have a tendency to propose H istorical Info for Smell deTection (HIST), an approach exploiting change history information to detect instances of five different code smells, particularly Divergent Amendment, Shotgun Surgery, Parallel Inheritance, Blob, and have Envy. We tend to evaluate HIST in two empirical studies. The first, conducted on twenty open source comes, aimed toward assessing the accuracy of HIST in detecting instances of the code smells mentioned higher than. The results indicate that the precision of HIST ranges between seventy two and 86 p.c, and its recall ranges between fifty eight and a hundred percent. Conjointly, results of the primary study indicate that HIST is ready to identify code smells that can't be identified by competitive approaches solely based mostly on code analysis of a single system’s snapshot. Then, we have a tendency to conducted a second study geared toward investigating to what extent the code smells detected by HIST (and by competitive code analysis techniques) mirror developers’ perception of poor design and implementation choices. We have a tendency to concerned 12 developers of 4 open source projects that recognized additional than 75 % of the code smell instances identified by HIST as actual design/implementation issues.