about: bad ideas


I have a lot of very weird ideas. Some are things like ``If there were giant carrots floating in the skies, instead of clouds, then carrots would be worshiped as gods.'' Even among the ones I bother to write down, about seven out of ten are obviously not worth developing. Experience has shown that of the remainder, about two-thirds don't work out for one reason or another. One idea in ten is a good one.

The important thing to remember is that at the beginning, all these ideas look equally bad. When I first got the idea to use (?{...}) to invoke a callback in a regex debugger, it seemed absurd, a weird hack of a little-used feature to do something that nobody was sure was worth doing in the first place.

I think if I had been asked on day 1 to estimate the maximum possible value of this idea, I would have supposed it might be good for an April Fool's prank of some sort. But it turned out, after we put in the proverbial 99% perspiration, that it was a workable idea, and probably even has some commercial value.

So lesson number one is that most non-obvious ideas look pretty strange the first time you see them, and you shouldn't send them away even if they have horns and snaggle teeth.


source: Rx: A Regex Debugger for Perl, by Mark-Jason Dominus.
keywords: idea
date: 12/31/2004