Re: [gnugo-devel] pattern-based reading spinoffs (27.7)

[Daniel Bump]

> Yes, I have no doubt that my scheme suffers!  However, it does
> gain a state over the other schemes because it doesn't use
> a PASS state.  PASS is implied if kom_pos == NO_MOVE.

That's a good idea ... I wonder if it is possible to cram
more information into kom_pos.

Actually it looks to me as if 9 bits would be large enough
for kom_pos if MAX_BOARD is at most 19 so BOARDSIZE <= 421 < 2^9.
This means that we could add one bit to komaster unless I'm 
missing something.

Dan