The three-evening rule

I have a projects folder holding twenty-three folders. Eight of them contain working code. The rest are a name, a two-line README and a promise to come back.

I used to think this was a discipline problem. Then I counted differently: if a project hasn't grabbed me within three evenings, it never will. First evening — interesting, second — understandable, third — either something alive appears, or it becomes clear I'm doing this out of stubbornness.

The rule is simple:

• Three evenings in a row, if there's time.
• If you don't want to open it on the third evening, close the folder and don't come back.
• If you do want to, keep going — even if it comes out badly.

Side effect: fewer abandoned projects. Not because I finish more, but because I stopped starting the ones that obviously won't reach a third evening.

The funny part is that the two most useful tools in my life grew out of exactly such three-evening sprints. Both were written in a hurry and both still work. Meanwhile the beautiful ideas with a month-long plan are still sitting in the folder.