What I think I know about testing

1. Convincing your organization that unit tests are a good investment is a 1999 problem. Everyone tests now, the conversation should be about testing effectively.
2. The goal of testing is not to produce a good design. At best, that’s a nice side-effect.
3. The goal of testing is not to document the business. At best, that’s a nice side-effect.
4. The goal of testing is to catch bugs and prevent regressions.
5. Prescribing TDD is like prescribing top-down programming or dark mode. If it works for you, fine; if it doesn’t, fine.
6. Not every layer of the code needs to be tested in isolation.
7. Hell, not every layer of the code needs to be tested.
8. The more tests are not the merrier.
9. It’s OK to skip tests on a prototype.
10. If it doesn’t have tests, it’s a prototype.
11. Coverage metrics can tell you what you missed. Targeting a specific coverage number is silly.
12. Integration tests are almost always preferable to unit tests.
13. Integration tests should be pushed as close as possible to end-to-end tests—without making them slow, brittle, or hard to set up.
14. Mocks are a bad smell.
15. The database should not be mocked.

16. Beyond helper functions, trying to remove duplication makes test code worse.

    
    

(I’m a proud grug-brained developer.)