On the Technical Debt Prioritization and Cost Estimation with SonarQube Tool

Abstract

Commonly, software developers are faced with situations to compromise internal quality to achieve short term goals, e.g. time-to-market. In software engineering, such compromises are described with Technical Debt (TD) concept. TD implies cost of additional rework—usually expressed as effort—and when the code is compromised, it is called code debt. One of the most popular tools for identifying TD items and estimating effort for solving them is SonarQube. However, there is still a need for empirical validations of remediation times that SonarQube estimates. The objective of this research is to validate the usefulness of the tool for novice developers in terms of understanding SonarQube’s categorization of criticality levels of TD issues and accuracy of estimated remediation times. We designed and conducted an empirical study with 185 students in a context of university software engineering courses in Finland and Serbia. The estimates provided by SonarQube are more pessimistic, i.e. only in 3% of cases the actual fixing time was higher that the estimated one. Our results also indicate that participants’ perception of the criticality levels of the TD items is misaligned with SonarQube’s classification and prioritization.