Re: [gnugo-devel] snapback, connect.tst 23, 39

[Daniel Bump]

Gunnar wrote:

> It has not. For the time being readconnect.c is to be considered as
> highly experimental code. I leave it to Tristan to decide whether he
> wants to try to use the tactical reading in reading.c together with
> his design for readconnect. The patch in the message referred above
> shows that the tactical reading in reading.c can be used with
> advantage together with the readconnect code at its current state, but
> since the "relevancy zone" concept has not been implemented yet it's
> not clear that this will remain the case.

One thing we might think about is implementing a more precise
version of the reading shadow in the incremental board which
could be used to compute Tristan's relevancy zones. This might
not be so easy. On the other hand, the benefits might be substantial.

Dan