Re: computer-go: Engineering (was: Most simple Go rules)

[Don Dailey <drd@xxxxxxxxxxxxxxxxxxxxx>]

   From: "Mark Boon" <tesuji@xxxxxxxxxxxxxxxxx>

   It seems we disagree on a large number of points. That number would be
   greatly reduced however if you'd read what I write a bit more carefully  and
   try to understand it instead of writing criticism on impulse.

Mark,

I actually think there is much  agreement here.  What  I see as really
happening here is this:

    1. Most of us agree that Japanese is broken.
    2. Some of us are more traditional and are willing to live with it.
    3. Some of us are anal about this and want a clean fix.


What I would  like to  see is  a protocol,  perhaps  the one  you have
invented,   that can accomodate  every  ruleset with some consistancy.

However I would  also like to see  agreement of Tromp/Taylor rules for
computer/computer tournament  play and make  an attempt to promote it,
with the   goal of it gaining  wide  acceptance in  tournaments.  That
probably won't happen if  it's not widely  endorsed by the programmers
themselves.

If there  is an agreement protocol  like yours, then it doesn't matter
if some ruleset is broken.

The  only  thing I  see  as being difficult   to deal with  is writing
software that  can correctly arbitrate 100% of  the time, using broken
rulesets.   I cannot build an  autotester  for instance, that  doesn't
take    some  liberties in  adjudicating    games,  or requiring human
intervention.

Now  that situation is   great for  humans,  who  have  an   excuse to
socialize and  bask in    traditionalism,  but  it's sloppy  from   an
engineering point of view.  Us  anal types are disturbed knowing  that
we  cannot write software   that   actually  works.  It's even    more
disturbing that this software is used in tournaments.

I agree with  you  that we  agree more  than we  disagree and are  not
always listening.


Don