Re: [gnugo-devel] experimental code

[Arend Bayer]

On Fri, 16 Aug 2002, Evan Berggren Daniel wrote:

> OK, so it looks to me like fixing owl inconsistencies requires a far
> greater understanding of the codebase than I have.  I'm happy to continue
> work on matcher_check (does it need a new name now?) if that is desired;
> would people have a use for replaying of sgf files or do we have the info
> needed?

It's definitely great that you have implemented this now. With regards
to understanding the codebase: Owl is probably the most difficult to
understand part of GNU Go, and it simply takes time to get to know it.
However, I would think that analyzing sgf-games (I'd try replaying games
from NNGS) might turn up a lot of easier to solve deficiencies of OWL,
e.g. missing basic simple patterns.

Arend