Author: Dann Corbit
Date: 09:04:15 03/26/99
Go up one level in this thread
On March 26, 1999 at 11:33:52, Gene Ward Smith wrote: >On March 25, 1999 at 20:06:33, Todd Durham wrote: > >>First, it would seem to me (and I confess I have only been involved in this for >>roughly a month) that the worst damage is to people who aren't yet involved in >>computer chess but who might gain an interest (as I did recently). > >I post from time to time on rec.games.chess.analysis, but posted by first >article to rec.games.chess.computer just a few days back. It generated >interesting and informed replies from, among others, Bob Hyatt. It was >a little bit unsettling to have this event happen so precipitously after. >While I am a mathematician rather than a computer person, I had been thinking >of looking at the code, but now that seems attended with difficulties. While >the peculiar atmosphere of rec.games.chess.computer probably served to >discourage some people in taking an interest in this subject, I suspect >retreating to a zone of privacy will discourage involvment by novices even >more. All in all, an unfortunate situation. My degree is also in mathematics (Numerical Analysis actually). If you want to learn to write chess programs, I would suggest this sequence anyway: 1. Get Tom Kerrigan's TSCP: http://ucsu.colorado.edu/~kerrigat/tscp13.zip Study that until you understand it. 2. Get Dusan Dobes' Phalanx: ftp://ftp.math.muni.cz/pub/math/people/Dobes/ Study that one also. 3. Then try Crafty. Crafty is a bit overwhelming to try and learn to write a chess program. But when you want to know how to do the trickiest sort of thing, that is a very good place to look. I think trying to learn from crafty first is a mistake. There is about 1.5 megabytes of source code for crafty. There is 220K bytes for Phalanx, and 40K for tscp. The difficulty of understanding is even more than a linear size relationship would suggest. I have found that looking at the code and the comments only helps a little. The postscript documents on chess programming that can be found on the net are even more helpful than the code at first. But when you really want to understand something, there is nothing like tracing through source with a debugger.
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.