Sean M. Burke wrote:
> I disagree with the implications made on this list
> ("Memory is cheap; time is expensive" being a quotably short one)
> that AI is necessarily always time-critical.
I didn't imply that AI is necessarily always time-critical.
"Necessarily" and "always" we can dispense with out of hand.
But "time-critical" is quite subjective. Clearly you can't
put a terrain recognition system in your cruise missiles
that takes minutes to do one recognition. Lots of problems
people try to attack in the real world take hours or days
to solve, on heavy iron. I don't know if that's "time-critical",
but time is money, so...
> A time difference of, say, 4x is bad
> if you're talking hours or days, but if the timescale is in single seconds
Of course! At that scale, it's far more important to optimize
*programmer* time, not *program* time.
That's why (/when) we use Perl, after all.
> When AI is technology (i.e., you
> developing something slick for everyone to use), speed is often a concern;
Yes, and that is not uncommon now. Certainly much more
common than it used to be. But also consider that a lot
of AI-as-technology is in-house stuff, with which
companies solve their own problems, or hire out their
problem-solving capability. E.g. GeneticProgramming.com.
--
John Porter
