We fixed the same bug twice, two commits apart
One commit says measure the problem instead of guessing at it. The one after says measure with somebody else's instrument, because our own timing was wrong. Same error, one layer down.
September 2026 · 2 min read
Two commits, two apart, in the same repository.
The earlier one reads: measure the site problem instead of guessing at it. The later one reads: measure with Google's tooling, because our own timing was wrong.
We had fixed the same mistake twice without noticing the first time that we had only fixed half of it.
The shape of it
The first commit stopped us guessing at the problem. We had been asserting that a prospect's site was slow. We replaced that with a measurement, felt pleased, and moved on.
But we were still guessing at the instrument. We had assumed our own timing was accurate enough to stand behind, which is exactly the same class of error one level down: an unexamined assumption sitting underneath the thing we had just carefully examined.