News
Ship the broken thing
A project sat at eighty percent for four months because I was waiting for it to be finished. I published it as-is. Two days later I knew exactly which twenty percent was missing, which is more than I knew before.
The rule I had been following was “don’t publish until it works”. It is good advice for software and terrible advice for a person working alone, because nobody else exists to tell me the thing works, and I had become the least qualified person to judge it.
What actually broke
- The navigation had three links pointing at pages I never created.
- The first screen took four seconds to appear on a cold load.
- Nothing crashed. Nothing lost data. Nobody was harmed.
What actually mattered
Two hours after publishing, the parts I had spent the previous month polishing were the parts nobody mentioned. The missing navigation was mentioned by three people before lunch. Effort and importance turned out to be unrelated axes, which is an uncomfortable thing to learn in public and a very cheap thing to learn.
Perfection is not a standard you meet. It is a delay you schedule.
The new rule
- If it opens without an error, it ships.
- The page says plainly what is not finished.
- Missing links are allowed. Broken promises are not.
Point three is the only one that took effort. A post that admits it is half-finished reads as honest; a post that quietly links nowhere reads as abandoned. Same state, different sentence, completely different impression.