Re: computer-go: new topics

[Darren Cook <darren@xxxxxxxxxxxxxxx>]

> The lure of perfect play is undeniable. However, at boardsize
> 5x5, it is also irrelevant.  Perhaps some people are enchanted by how
> successful computer-chess has been using brute-force-based methods.  Perhaps
> some are enchanted by the exponential growth of dirt-cheap computational
> resources.  Either way, the discussion at hand is in no way furthering the
> progress of computer-Go.  Let's not kid ourselves, this is a combinatorial
> math problem, not a computer-Go problem.

"Computer-Go" includes both the commercial programs that play on 19x19
and solving go on smaller boards. The latter tend to be of more interest
to academics, but can still provide some useful discussion for
larger-board programmers. E.g. understanding how to store compactly and
look-up quickly in a pattern database. And how ko affects what you
should store.

Darren