Treat Your Code as a Crime Scene
summary
Uses git commit history and CodeCity-style maps to find hotspots (complex + frequently changed) in a 400,000‑line codebase where 4% of files account for 72% of defects across 89 programmers. Adds temporal coupling to reveal files that change together (copy/paste clones or hidden dependencies) and uses fractal figures to map which developers or teams own code. Shows complexity‑trend graphs to predict decay, flags abandoned modules, and mentions custom open‑source tools plus D3/Processing visualizations.
AI-written summary · about
chapters (15)
- 0:00Start
- 0:50Understanding Code
- 2:21Cyclomatic Complexity
- 9:24Profiling the Ripper
- 12:41A geography of code
- 13:55Spatial Movement in Code
- 14:39Geographical profiling of Code
- 15:08Case Study
- 16:52Hotspots and Defects
- 17:50Evolution of a Program
- 20:02Complexity Trend
- 35:40Identify Main developers
- 38:15Social Networks in Code
- 39:31Loss of Knowledge
- 40:20Code 3 Forensics
related talks
49mTreat Your Code as a Crime Scene
40mTreat Your Code as a Crime Scene
31mMaking sense of frontend code with forensic techniques
39m7 Secrets of Maintainable Codebases
38mBeyond Code: Repository Mining with Clojure
41mA Short History of All People ...
38mCode reading
29mTaming A Monster With Git Driven Data