Quality

Reviewed: 2025-10-10

Prof. Jürgen Börstler

Questions regarding the research presented on this page? Contact Prof. Jürgen Börstler.

SERL investigates software quality where it matters - in code reviews, CI pipelines, testing assets, defect handling, data, and product documentation. Our work on modern code review explains how information spreads among reviewers, how to prioritize requests, and how to make comments and commit context more actionable. Practical guidance supports the review of GUI-testing artifacts, while experience reports identify adoption pitfalls. NLP-oriented analyses highlight what to watch for when mining review comments and how to identify prevalent quality issues from reviewer feedback.

A second stream clarifies what to measure and why. Catalogues of source-code and GUI-testing metrics and evidence on links to external quality characteristics help teams choose indicators to track maintainability and reliability. For tests, our work proposes improved smell detection and details which test-suite attributes matter most in practice. In our work, we also position CI as a vehicle for evaluating non-functional requirements and for visualizing quality evaluation.

Quality is also treated as an organizational outcome. Results connect internal quality and customer value, relate technical debt to lead time, and show how ownership, clones, and backward compatibility policies shape technical debt. Additional contributions include data-smell catalogues for AI-intensive systems, requirements/documentation debt, and developer views on “clean code.”

Current and Future Work

Trends point to context-aware code reviews that blend reviewer guidance, diffusion insights, and NLP summaries to spotlight risk and recurring quality issues. Measurement work is converging into lean dashboards aligned with maintainability/reliability and CI-based evaluation of non-functional qualities. Industrial ML for bug triage will emphasize explainability and adoption hurdles. Debt management will integrate ownership/clones analytics and policy costs.

Together, these insights help teams make quality visible, auditable, and improvable within everyday workflows - reducing waste and accelerating value delivery.

Important context

This text was generated by AI and edited by humans. It is based on SERL's research publications between January 2020 and September 2025. For technical questions, please contact Dr. Michael Unterkalmsteiner.