Author: Dann Corbit
Date: 14:25:02 08/20/02
Go up one level in this thread
Marcel responds (quoted with permission):
I have not had a subscription to that webforum since 1998 as I was not really
impressed what went on over there. I stopped reading such snake-pits a long time
ago and have never felt any better :) I have to add that this 'CCC' is not the
primary audience for this work. Also, the work is dual: for a big deal,
page-wise, it is a lengthy introduction to the field ('boring repeat' if you
like). Unfortunately, that is needed to get the intended audience up to speed,
which you may or may not like in such a work. It has some unintended side-effect
that it could act as an introduction text, which is nice, but not the purpose.
The second and, to me, the most important aspect is that presents quite a few
new and unpublished ideas on all areas, albeit somewhat mixed with the whole
story: It gives quantitative evidence for improved hash table usage for n-probe
(n>2), it casts serious, quantative, doubt on the effect of the counter-move
technique in combination with domain-smart (SEE) move ordering, it solves the
lazy-evaluation inconsistency problem, rebutes the hash-miss approximation
models that have been pestering academic publications before (even in PhD's). At
the same time it is an engineering work that, among others, demonstrates a new
algorithm that solves many hard pattern recognition problems in software
evaluators. There is more in it that I haven't even bothered to mention here.
For what it is worth, until now, all of the chess programmers I know who have
read it responded quite well.
This next part is edited by me [DRC]:
[snip]
"...my supervisor is in [snip] software engineering with a professional interest
in puzzles and games. Second, an ICCA board member and professor of the
Maastricht University Games group was in the committee."
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.