Notes on craft

Some things I've started doing on purpose:

Ship the ugly version first. Polish is the easy part. The hard part is finding out whether the thing was worth making at all. The faster you find out, the cheaper the answer is.

Write the README before the code. If I can't explain what I'm building in a paragraph, the project isn't ready. Sometimes I find out it never was.

Limit the scope, not the ambition. The cure for a too-big project is rarely a smaller dream. It's a smaller first slice of the same dream.

Read your own work tomorrow. I almost never publish on the day I finish writing. The 24-hour version of me is a much better editor than the version that just typed :wq.

Keep a "next time" file. Every time I get bitten by something — a config gotcha, a wrong abstraction, a habit I want to break — it goes in the file. I read it before starting anything new. The list gets longer; the same mistakes get rarer.

None of this is original. All of it is practiced.