Author: Bruce Moreland
Date: 10:44:37 11/14/99
Go up one level in this thread
On November 14, 1999 at 06:46:10, Bella Freud wrote: >I might also add that his "sharing" leads to underperformance as his competitors >steal his ideas (parallel, bitboards) and return him nothing but abuse. >Therefore any competent programmer will always be able to produce an end-product >that is better than his Crafty. > >In my opinion the greatest attacks on Mr Hyatt come from commercial and wannabee >commercial programmers who have benefited mostly from Mr Hyatt's code and ideas, >but now want to distance themselves from the man himself in order to pretend to >themselves that they did it "alone". These are interesting ideas but I don't think that they bear up. I think that Bob has made a large contribution with his articles on Cray Blitz, and techniques described in these are probably incorporated into many programs. This is because these articles were early, and anyone who wanted to start a chess program in the late '80s or early '90s encountered them if they did any sort of literature search at all. The work with Crafty was not really popularlized until 1996 or so. Most of the strong commercials were strong commercials by 1996, and there is no evidence that I can see that any quantum jumps in strength came from Crafty, unless you count the idea that a chess program should be a null-mover with end-point eval and a chopped quiescent search, which is something that several of the ICC'ers evolved towards simultaneously, and the most available example is Crafty. There are two ways you can build a strong program with help from Crafty. 1) Eat the whole thing, the full-body transplant method. In this case you wind up with a great wad of Bob and perhaps a few ideas of your own on top, which may have some influence, but may have no influence. This is essentially a renamed Crafty, and as such is not an original work, it's a direct copy. 2) Read Crafty source as if it were an exceptionally long and well-written article. You can get information about search mechanics, extensions, null-move, evaluation terms, etc., and you can get some specific stuff. This can be of much use, but it doesn't immediately follow that every new program will be as good as Crafty based upon this. Anyone who views computer computer chess as a personal challenge will still write their own program, will still have the traditional two-ply search to start with, and will still suck for several months if not longer. I think that Bob is an important guy, but it is not like if you read Bob's stuff all of the work has been done for you. bruce
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.