Comments on: User Friendly Heuristics

By: God of Biscuits

God of Biscuits — Sun, 23 Aug 2009 00:31:00 +0000

Coming up the other way, from molecular biology up, software development practices are following quite closely evolution, even abiogenesis.

Evolution doesn’t pursue perfection; it doesn’t “pursue” anything, really. And every result it produces that can be observed and interpreted as an improvement happens because of generational mutation.

We stand still in one generation; our iteration happens in development and i think we forget that sometimes we get the fitness function wrong (“graceful degredation” vs “perfection”) and end up iterating/evolving in the wrong direction.

By: John Gallagher

John Gallagher — Sat, 22 Aug 2009 13:08:41 +0000

This is a very relevant post for me right now – I’m trying to solve a problem that I’ve not seen anyone do yet. And it’s a really hard problem. I just need to accept that version 1 will not be perfect, but focus on what I can do that will provide most value to the user.

Thanks for your post, Daniel.

By: Christophe Pettus

Christophe Pettus — Fri, 21 Aug 2009 18:29:06 +0000

And, in an example of how this kind of thing can be insanely, amazingly subtle… Wil’s sample code doesn’t work. Why? ISBNs can contain the letter X (but only as the last character) as well as [0-9]. I forget that all the time, too, and it’s another example of how you have to keep tweaking, tweaking, tweaking…