Re: [gnugo-devel] more joseki preparations

[Arend Bayer]

Paul wrote:

> Dan wrote:
> > Paul wrote:
> > > The new matcher itself is written but most certainly won't work
> > > for i haven't even tested it and there must be many bugs. it is
(...)
> > Could you explain what your plans are for this?
>
> currently my plans are to finish it :). it is supposed to become a
> very fast pattern matcher (i hope so ;) specifically designed for
> joseki database. the main intention is to use it with joseki reader
> some day - an idea by Evan and Inge as far as i remember. it (the
> reader, not the matcher) should be able valuate several available
> joseki continuations taking into account moyo and and general
> influence of distant stones - e.g. select the side on which to

Can you maybe explain a little more? What you describe would be possible
with the old fashioned pattern matcher, too. Will it have transposition
tables?
Also it would be good if you added comments with your next patch.
(A descriptive comment at the top of each non-trivial function is
usually enough.)

Arend