On Jul 20, 2007, at 7:23 AM, Jason House wrote:
Thanks for the documentation. I have a few questions.
Looking at only the four neighbors to detect eye-like points seems
like it could leave many false eyes and allow captures of dangling
chains. Is there any mechanism to avoid this problem in the play
of the bot?
It does also look at the diagonals; see Board.isEyelike(). I'll note
this in the next version of the document.
I'll have to think more about the multi-stone suicide thing.
There's definitely a few situations where that can change the life
or death status of a group. How much of a speed difference do you
get from the suicide thing?
I haven't run those experiments myself, trusting to Lew's
optimizations. Perhaps he can weigh in on this.
The documentation implies that you're not using a disjoint set for
chain id tracking. I thought this was one of the large speed gains
in libego.
I'm using the same data structure as Lew. Each stone knows its chain
ID number, which can be used to look up it pseudoliberty count. I'll
hazard a guess that this is faster than traveling up a disjoint set
tree, even with path compression.
Peter Drake
http://www.lclark.edu/~drake/
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
