A survey on developer’s intention upon refactoring: Assessing refactoringminer’s efficacy.

Abstract

Refactoring is an essential practice in software development, as it allows developers to improve design, readability, and maintainabil ity. In this context, analysis tools such as RefactoringMiner intends to provide an arguably precise classification of refactoring types. However, there is a concern regarding the alignment between the refactorings identified by these tools and the developer’s intention towards those changes. For example, the tool may fail to detect the developer’s intention of doing a Pull-up Method refactoring. In this work, we are going to address this issue. Concretely, we will conduct a survey with over 200 experienced Java developers that contribute to highly rated and active open source repositories. Those developers will be surveyed regarding refactorings detected on specific commits they authored. The goal is to complement the evaluation provided by RefactoringMiner, asking developers whether they recognize and had the intention of performing the refactoring types detected by the tool rather than assuming the output as correct. Another goal is to assess whether the tool failed to detect refactorings performed by the developers.