Author: José Carlos
Date: 16:45:55 05/05/02
Go up one level in this thread
On May 05, 2002 at 17:17:56, David Dory wrote: >On May 05, 2002 at 15:38:44, Fernando Villegas wrote: > >>All those guys I mention above these lines were, in one moment, authors of good >>programs. And all those guys are by now retired from chess programming. >>Certainly there are probable other names I do not recall. What happened to them? > >Life happened, I guess. > >Some of them, I'm sure, had personal reasons for retiring, not related to chess. >In the Spracklins case, the "team" got divorced, and their project with Saitek >was not stunningly profitable. > >Yes, I believe there is a creativity factor at work. In the early days of chess >programming, it was so exciting just to see a computer play chess at all. A new >idea could generate a whole re-write, and net you a huge improvement in playing >strength. Consider the change when Northwestern's CHESS went from ver. 4.2 to >4.5. Slate & Atkins added something new - a transposition table! This was big >and exciting! Would a chess program become unbeatable with this stunning >breakthrough? > >Now, our software improvements are all rather normal, hardly edge-of-your-seat, >stuff. Software improvements have been hard to come by for many years (10-15). >As the chess program matures, you wind up working just as hard, for (generally), >smaller and smaller returns. You can't fight this - it's just the way it is. You >already eliminated the big bottlenecks in the first few years, only the smaller >one's are left to work on. As you say, the programmers aren't working with big >numbers now, it's little "decimal places". This may not be what attracted these >people to chess programming in the first place. SMP can be a big jump when multiprocessor machines get big number of processors. Yes, everyone can walk that way, but only some people have the time and talent to make it work at highest level (same thing happened at the first stages of transposition tables, null move, ...). I believe (and hope) there'll will come new possibilities like SMP that will keep challeging chess programmers. Computer science advances so fast... José C. >I think everyone would like to give their chess program one big push, at least >once, before they retire from working on it. > >I think Bob and Ed are exceptions, and exceptional. You'll know when Christophe >no longer feels creative - he'll start saying that one's creativity can only >last for just so long! :-) > >Hopefully, that won't happen for creative or economic reasons. > >Good post - good jaunt down memory lane. > >David > >P.S. Bryant's Colossus was a checker's program. Played against Chinook several >times. Did it play chess also?
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.