Hotspots, Change Coupling, and Crime Scene Analysis

Your Code as a Crime Scene: The Psychology Behind Software Quality - Adam Tornhill

Topic 5 of 10

Published Dec 1, 2025 · Clip 24:33–33:27 (8:55)

Adam explains the core idea that not all code is equally important: change frequency follows a steep power-law curve where 1–2% of files account for the majority of development work. These 'hotspots' are where quality improvement pays off, while rarely-touched code can tolerate technical debt. He critiques static analysis tools like SonarQube, which produce thousands of unprioritized issues and drown important hotspots. He also covers change coupling—revealed via Git history—which maps how files change together, aiding onboarding, architecture refactorings, and diagnosing costly god classes via the hotspot X-ray technique.