Don't Hang Me Out To DRY
summary
Claims that 100% test coverage proves safety collide with concrete failures: a mercury-in-retrograde bug, flaky probabilistic tests, and an opaque access‑key generator that was over‑abstracted. Tools actually used include Ruby, Rust + Helix interop and mutation testing mentions. The conference codebase shows FizzBuzz‑as‑a‑Service built in Rust, surprising maintenance costs and unclear performance gains, and a DRY abstraction that made changes brittle rather than easier to evolve.
AI-written summary · about
chapters (13)
- 0:00Intro
- 1:21MAKE IT RIGHT
- 5:17Code Coverage
- 6:19Coverage is insufficient
- 6:57Pair Programming
- 7:07Mutation Testing
- 10:29Vanity metric?
- 10:54Coverage is a signal
- 13:35Total Cost of Ownership
- 17:32Descriptive And Meaningful Phrases
- 23:07Flexibility
- 24:47Performant Code
- 29:45Weigh observed behavior & maintenance
related talks
54mSeeing Like a Programmer - Resiliency, Limits, And Moral Hazards In SE
17mIf You’re Happy and You Know It (Inside the Mind of a Developer)
40mUnaccustomed As I Am to Public Speaking
40mWorst. Ideas. Ever.
27mYellow Brick Code: Lessons from The Wizard of Oz on Enhancing DevEx
01h 02mZen and the Art of Code Lifecycle Maintenance
31mYour Tests Aren’t Flaky, You Are!
15mWhy TDD Is Crap