thought out what
he wants. In some contexts it would make sense to
completely get the
goals worked out and the algorithms proven before starting
to code. That makes sense when the coding is the hard
part. You code less when you can
do it right the first time. But he's groping toward a hard
concept.
When the idea is harder than the code, it makes sense to
try things out
and see them. He'll throw some code away, but it's a check
on his
thinking. For hard problems it helps to have an
interactive computer to
test things on. (And if you're working under a deadline,
it makes sense
never to accept a contract that includes a hard problem.
Get the specs completely straight, code it once, and you're
done. Reject all contracts where the specs aren't firm.
Unless you can't get enough contracts that meet those
criteria, and have to accept others to stay afloat.)
>Just what *is* your targeted audience?
I wasn't particularly noticing that. Somebody asked a
question and I answered. Stimulus/response.
I'm not clear why Mentifex is stressing Forth, I'd have
thought he'd be
most interested in his AI idea. Forth is a very good tool
for the task
he's attempting, but it's just a tool. If somebody proves
Hrudinski's Conjecture and publishes it, I doubt they'll
make a big deal out of
whether they used LaTex or Postscript or whatever to do the
layout. If
they used a calculator I doubt they'd make a big deal
whether it was HP
or TI. Or Mathematica or Maple. If they designed their
own special processor to solve the problem I doubt they'd
sing the praises of Verilog
or VHDL.