When Nobody Owns Quality: Making Testing Responsibility Explicit in Open Source Projects
Talkat Open Source Summit India 2026 · on Wednesday, 2026-06-17· at 13:40· Lotus Ballroom 3 (Level 3)·
summary
Open source projects often have CI and tests but no clear enforcement, so quality stalls until governance changes. Evidence: Kafka drops from 90% failing main builds and 0.1% PRs fully green to near-normal after explicit policies; Linux reduces ignored regressions with Ragebot regression tracker; Chromium prevents merges when CI breaks with Chromium Sheriff blocks the tree. Recommended actions are concrete: write testing ownership into governance, set enforced release gates, designate CI health stewards, and track flaky-test debt.
AI-written summary · about
chapters (11)
- 0:00Open source quality challenges
- 1:52Quality unowned in open source
- 7:15Testing without consequence
- 14:04Linux Ragebot for regression tracking
- 18:25Authority without enforcement
- 22:51Chromium Sheriff model for CI
- 25:49Ownership gaps at team boundaries
- 29:44Signal consequence gap
- 32:15Governance improved quality
- 36:42Kafka governance turnaround
- 38:08Q&A
related talks
46mProject Culture and Team Responsibility for Quality
23mGood Governance Practices for Healthy Open Source Projects
35mWhy Software Engineers Should Make Quality a Priority
31mOperationalizing Open Source Projects: When the Rubber Meets the Road
49mCommunities Over Code: How to Build a Successful Project
42mImproving Testing Quality by Changing Culture
23mTesting - Best Practices in Open Source World
44mIs this Open Source Project Healthy or Lifeless?