Re: computer-go: About Viking

[Tristan Cazenave <cazenave@xxxxxxxxxxxxxxxxx>]

Eric Boesch wrote:

> >I almost forgot my main principle:
> >There should be no compromises in how the program represents
> >things. Either a chain of stones is connected or it is not, and
> >the program should know exactly why.
>
> same principle -- "the only direct way to prevent the capture is to play on
> one of the group's liberties, or to play on a liberty of a potential
> capturing stone".  Indirect attacks and defenses are just chains of direct
> attacks and defenses.  The "knowing exactly why" part, then, involves
> remembering which stones and liberties had an indirect effect on your final
> conclusion.
>

I also use exact tactical computation. Surprisingly, they

can be faster than non-exact ones. Finding the intersections

that are potentially involved in a tactical computation, when

the goal is 5 or 7 plies way can be very tricky.

I am quite interested in developing automatic tools

to find such knowledge. Any help about that topic would be

appreciated. OF course, one can use a kind of dilatation

operator on the strings so as to be conservative, but this

involves too much intersections. The ideal would be a compromis

between simplicity (dilatation) and accuracy (only the exact

relevent intersections).

--
Tristan Cazenave
Universite Paris 8, Departement Informatique, Labo IA
2, rue de la Liberte 93526 Saint-Denis Cedex - France
http://www.ai.univ-paris8.fr/~cazenave/