Road to Zero Lint Failures: Tackling Code Quality Challenges at Scale
summary
Claim: lint failures drop from a ~7,000 peak to near zero by combining policy, tooling and people. Built a lint failure visualizer, a nightly checkup cron job that converts ESLint output to SARIF and a weekly scorecard email; enforced single owner per file, a lint proposal process and a volunteer task force. Result: ~6,000 failures removed over ~13 months with 55 engineers, 80 rule/dependency changes and 45 contributors; first 80% clears quickly, the last 20% requires hands-on effort and active incentives.
AI-written summary · about
chapters (34)
- 0:00intro
- 0:22preamble
- 0:33overview
- 0:52why lint rules?
- 2:01automated
- 3:24prettier
- 3:42benefits
- 4:35example: no-unsafe-optional-chaining
- 6:47expand the "well lit path"
- 7:05let's add all the lint rules!
- 7:17code quality at scale
- 9:02problem
- 9:34campsite analogy
- 10:08how to limit even more lint failures
- 12:11campsite analogy (revisited)
- 13:08how to limit even more lint failures (revisited)
- 14:44lint rule process: all existing failutres must be resolved
- 16:41how to fix it: road to zero lint failures
- 16:57incentives, incentives, incentives!
- 17:26road to zero lint failures
- 17:56how we ran it
- 19:10tooling for engineers
- 19:41single owner per file
- 20:10lint failure visualizer
- 21:30checkup
- 22:43collect data from sarif and match by team
- 23:38weekly scorecard
- 25:47dealing with the last mile
- 26:38road to zero journey
- 27:20sustainability is key
- 28:20summary
- 30:45key results and next steps
- 31:40road to zero lint failures stats
- 33:09thank you
