Re: [Question] Re: TD & Intelligence

[Mousheng Xu <moushengxu@xxxxxxxxxxxxxxxx>]

At 12:22 AM 5/7/99 +0200, Joan Pons Semelis wrote:
>I'm afraid this 2 questions might be a little discouraging.
>But of course feasibility depends on the answer.
>I think nevertheless that un unfeasible approach is not that bad:
>once you have it you can think of a way to make it or prove that
>that was indeed the only way, so the problem had no solution.
>The program I'm writing is basically what I had in mind 10 years
>ago, I've passed this time trying to find a faster approach or
>faster computers, the second happened first :-)
>
>Joan

Joan, 

	I agree that we could entertain ourselves just by testing things out. The
go program based on NNW or Genetics/Evoluation sounds very hopeful, at
least to me 10 years when the Evolutionary Strategy (ES) came to my mind.
But I have not even tried it because it's practically proven to be very
slow. I have some other ideas in mind similar to ES. If I plan to work on
it or NNW or ES, I will be much more comfortable to have a brief idea how
fast the speed is and how big my hard drive needs to be.
	Even a simple, straight, "silly" pattern matching go has taken 1.5 years
of my spare time. Think about it, you could write two small-medium
comercial products within that time of period.
	Another main reason I asked the two questions is that I do want to know
some answers. :)
	Thanks.

-- Mousheng