An Experimental Investigation on the Innate Relationship between Quality and Refactoring

Previous research has investigated the reasons behind refactoring actions performed by developers, and proposed methods and tools to recommend refactoring actions based on quality metric profiles, or on the presence of poor design and implementation choices, i.e., code smells. Nevertheless, the existing literature lacks of observations about the relations between metrics/code smells and refactoring actions performed by developers. In otherwords, are unknown the characteristics of code artifacts pushing developers to refactor them. This paper aims at bridging this gap by analyzing which code characteristics triggers the application by developers of 52 different types of refactoring operations. To this aim, we mined the evolution history of three Java open source projects investigating if developers perception of code artifacts needing refactoring is inline with indications provided by quality metrics, change-proneness, and code smells presence. Moreover, we verify if developers are able to remove code smells from the system through refactoring. Results indicate that, more often than not, the developers' perception of which class needs to be refactored is not captured by quality metrics. Code artifacts having a high change-proneness attract more refactoring operations focused on improving code readability. Finally, 42% of refactoring operations are performed by developers on code smells. However, only 7% of the performed operations actually remove the code smells from the affected class.