Author: Georg v. Zimmermann
Date: 12:06:20 11/10/00
Go up one level in this thread
Hi, When I started chess programming I did that because I had lots of ideas I wanted to test and was sure I could make a real strong program with it ... over time I realized that most of those ideas don't work, some work but have major disadvantages, some would work but are very tough to get to work right. The best cure from strange ideas is try to make them work. Pick up a book on C programming and start today. Programming is not hard, even I can do it (a bit). This idea you mention is especially hard to program. And the problem is not how to use the computer memory you seem to focus on, it doesn't really matter whether your program + its knowledge libraries (to call them that way ) is 100 kb big in memory or 50 MB. (who cares about 10-500% speedup anyway). The problem is that it is hard to write the routines that tell the program whether it is time to use the special attack knowledge or not (for example). Also keep in mind that chess is not that simple that in one position there is one plan to follow. Often, for example, an attack can work because by using threats on the kingside and attacking a weak pawn on the queenside at the same point the defenders get too much to do. Which of the modules should be loaded now ? Both ? And who decides which one to follow ? And how about all computation overhead ? One last thing: you shouldn't complain about "silence" when you get 2/3 or more answers within hours. Many of my posts I didn't get a reply to at all. Best regards, Georg :) >Hi chap: >In the last years I have posted here almost exactly the very same idea, twice or >more times, even using the same word as you: modules. And the answer has been >always the same: it cannot be done. Or it has been partially done. But more >frequently the answer has been none. Zero. Silence. >As you, I am convinced that IT IS the idea. In one of those post I argued that >that is what a good player does: first to see what is relevant, then go to >calculations with specific criteria. If an attack is the core, then you think in >terms of sacrifices, etc. If not, then another modules comes to the task. >Of course there is another and more satisfactory alternative than you and me >could consider: we made bingo and half a dozen programmers, without saying >nothing, are working on it. >I even has some idea that in fact one program is following that path. Oh, not >because of me or you, but probably because the idea is in the air. Latent new >paradigm somebody could say.. :-) >Fernando
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.