7 Secrets of Maintainable Codebases
summary
Argues that Version Control behavioral data surfaces the real maintenance targets: a power‑law of change concentrates most work in 3–6% of files, hotspot analysis builds a probability surface from commits (one case used 2,400 commits) and highlights tiny yet costly files in huge codebases (Roslyn ~4M LOC, some test files ~7,000 LOC), shows hotspots predict defects, uses temporal coupling to find files or services that change together (example: 89% co‑changes due to copy‑paste), and combines code with social maps like hotspot visualization and fractal figures to expose abandoned areas after key developers leave.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 0:36Let's Turn The Microscope
- 1:27Behavioural Data
- 2:10The challenges of Scale
- 4:59Change Distribution of Files
- 6:04Focus on the Code that Matters
- 12:34Hotspots in Roslyn
- 15:37Prioritize with Hotspots
- 19:16Where's the Gorilla in your Code?
- 24:05The Cost of Surprise
- 25:50Physical Coupling
- 26:37Logical Coupling
- 28:31Analyzing Architectures
- 30:30A Social View of Clojure
- 31:35Excess Parallel Work?
- 32:09Why Hotspots stay where they are
- 34:17A Main Developer leaves Scala
related talks
40mTreat Your Code as a Crime Scene
49mTreat Your Code as a Crime Scene
47mTreat Your Code as a Crime Scene
39mArch. App: Prioritizing Technical Debt as if Time and Money Matters
32mPrioritizing Technical Debt as if Time and Money Matters
20mPrioritizing Technical Debt as if Time and Money Matters
50mPsychotherapy for Code
41mHow to write maintainable code