An extensible, feature-based framework for fine-grained code quality assessment

Abstract

Context:

Assessing code quality is essential for maintaining and evolving software systems. While traditional tools like SonarQube and Snyk offer valuable insights at the application level, they lack support for feature-specific analysis, making it difficult to understand how quality issues are distributed across the functional structure of a system.

Objectives:

This paper introduces InsightMapper, a novel approach designed to bridge this gap by enabling feature-oriented quality analysis. The goal is to assess and compare the quality of individual features, identify feature-level hotspots, and support strategic maintenance decisions.

Methods:

InsightMapper leverages existing feature location techniques to project quality analysis results onto feature implementations. We evaluate the approach on three case studies, including a recognized benchmark in the feature location domain. These evaluations demonstrate InsightMapper’s ability to perform fine-grained, feature-oriented code quality assessment using results from SonarQube and Snyk.

Results:

The study shows that InsightMapper effectively reveals how quality issues are distributed across features, uncovers features with disproportionate technical debt, and supports prioritization strategies grounded in functional relevance. The approach also enables the computation of feature-level quality scores, facilitating comparisons between analyzers and across features.

Conclusion:

InsightMapper offers an extensible and practical solution for feature-oriented quality assessment. By projecting analysis results onto the feature of applications, it enhances the interpretability of quality data and paves the way for more targeted maintenance and evolution strategies.