Debug like a scientist
summary
Claims a repeatable method: a loop of observation hypotheses and experiments applied to three cases. Reproduces a race by interleaving the business logic and the config hook locally and fixes it by moving the hook. Unblocks hundreds of developers by stopping noisy logging that avoids a Docker Compose freeze. Resolves 502s every 30 minutes by clearing a boot cache that delays passenger startup. Emphasises notes shared across timezones and parallel hypothesis work. observe hypothesis experiment reproduce race by interleaving logic
AI-written summary · about
chapters (8)
- 0:00The frustration of debugging
- 2:56Adopt a scientific approach
- 6:04Case study: Race condition
- 10:48Case study: CI infrastructure
- 14:53Case study: System outage
- 18:21Scaling and coaching
- 22:14The debugging pyramid
- 23:28Knowledge retention
related talks
43mDebug Like a Scientist!
33mGetting Unstuck: Using the Scientific Method for Debugging
36mFixing Flaky Tests Like a Detective
29mFilling the Knowledge Gap: Debugging Edition
37mDebugging Without Borders
46mBlink and You’ll Miss It: Stories From Discovering Most Annoying Bugs and Test Flakes
29mBlink and you’ll miss it: Stories from discovering most annoying bugs and test flakes
10mBeyond “try harder”: Effective strategies to tackle bugs