Investigating the Relationship between Technical Debt Repayment in Issue Tracking System and Refactoring

Abstract

Background: Technical debt refers to design or implementation choices that can lead to increased maintenance costs in the future. The accumulation of technical debt not only raises maintenance costs but also hinders software evolution, making its early removal (i.e., repayment) desirable. In modern software development, TD-Issues exist as a means of managing technical debt through issue tracking systems. Although prior research has analyzed code changes related to technical-debt repayment within TD-Issues, no studies have examined whether code changes occur at all as part of TD-Issue resolution. Objective: This study investigates the relationship between technical-debt repayment and refactoring in TD-Issues, examining the extent to which refactoring is involved in the resolution of TD-Issues and clarifying its characteristics. Method: Using RefactoringMiner, we analyze commits linked to TD-Issues to determine the proportion that include refactoring operations and to identify the most frequently used refactoring operations. Results: RQ1) We found that 59.6% of technical-debt–repayment commits performed to resolve TD-Issues include refactoring, and over half of these commits combine multiple refactoring operations. RQ2) The refactoring operations applied tend to be fine-grained changes, such as data-type modifications and renaming. RQ3) When refactoring is performed, it is common for developers to simultaneously apply multiple data-type–changing operations involving variables, return values, and other program elements. Future Work: To support practitioners in managing technical debt, future work may identify concrete characteristics of technical debt by classifying types of technical debt and analyzing repayment patterns associated with each type.