RENAS: Prioritizing Co-Renaming Opportunities of Identifiers

TL;DR

RENAS is a technique that improves identifier renaming in Java by prioritizing related identifiers based on their relationships and vocabulary similarity, aiding developers in more accurate and comprehensive refactoring.

Contribution

RENAS introduces a novel method to identify and prioritize related identifiers for simultaneous renaming, enhancing refactoring accuracy in Java applications.

Findings

RENAS improved F1-measure by over 0.11 compared to existing methods.

The approach effectively identifies related identifiers based on relationships and vocabulary.

Evaluation on real change histories validated RENAS's effectiveness.

Abstract

Renaming identifiers in source code is a common refactoring task in software development. When renaming an identifier, other identifiers containing words with the same naming intention related to the renaming should be renamed simultaneously. However, identifying these related identifiers can be challenging. This study introduces a technique called RENAS, which identifies and recommends related identifiers that should be renamed simultaneously in Java applications. RENAS determines priority scores for renaming candidates based on the relationships and similarities among identifiers. Since identifiers that have a relationship and/or have similar vocabulary in the source code are often renamed together, their priority scores are determined based on these factors. Identifiers with higher priority are recommended to be renamed together. Through an evaluation involving real renaming instances extracted from change histories and validated manually, RENAS demonstrated an improvement in the F1-measure by more than 0.11 compared with existing renaming recommendation approaches.