Re: [computer-go] Spiral pattern matching and data structures

[Paul Pogonyshev <pogonyshev@xxxxxxx>]

Anders Kierulf wrote:
> > None except it turned out to be faster :) The reason depends on our
> > pattern database: There are many patterns like this:
> >
> > ..OO             reverse sente sagari
> > .OYX
> > .*..
> > ....
> > ----
> >
> > that only match somewhere at the edge, and the diagonal spiral tends
> > to find out faster whether the anchor ('Y') has the right distance to
> > the edge.
>
> Why not test the distance of Y to the edge before even starting to check
> whether all the points in the spiral match?

Because with DFA you match all patterns at once.  You cannot filter out
patterns in advance, you just get a list of matching patterns.  However,
with diagonal spiral you usually get this list faster, because non-matching
at-the-edge patterns typically not match at smaller depth.

Paul
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/