Meet the Social Side of Your Architecture
summary
Argues that version-control data reveals hidden organizational faults using fractal figures and a normalized 0–1 team-coupling metric to detect modules with excess coordination needs. Shows a case where adding developers (fourfold) to hit a deadline turns a one-year project into a two-year, higher-cost, lower-quality outcome. Uses change-coupling graphs where changing one service predicts edits in seven others, and recommends small team-sized services, package-by-component structuring, and tooling like CodeScene (and earlier codemaat) to measure and guide refactorings.
AI-written summary · about
chapters (5)
- 0:00Start
- 10:17Mental Models for Code
- 18:16VCS - A Behavioral Data Source
- 28:06The Two Forms Of Accidental Complexity
- 28:54Change Coupling
