Author: Sune Fischer
Date: 00:31:40 05/22/02
Go up one level in this thread
On May 21, 2002 at 19:20:14, Rolf Tueschen wrote: >On May 21, 2002 at 10:34:34, Sune Fischer wrote: > >>On May 20, 2002 at 18:14:30, Rolf Tueschen wrote: >>>>>Quotation from FIDE rules: >>>>>========================== >>>>>(a) >>>>>During play, the players are forbidden to make use of hand-written, printed or >>>>>otherwise recorded matter, or to analyse the game on another chessboard. They >>>>>are also forbidden to have recourse to the advice of a third party, whether >>>>>solicited or not. >>>>>[The only possible exception is that a player in a team competition may be >>>>>allowed to ask his captain "Should I accept his offer of a draw?" or "Does the >>>>>team need me to play for a win?". The captain or acting-captain must limit his >>>>>reply to an immediate "Yes", "No", or "It's up to you", without supplying his >>>>>answer after a detailed analysis of the position, and without making his answer >>>>>emphatic in any way. This captain, like all his players, is not allowed to >>>>>receive opinions, from any source, on the states of play of any games still in >>>>>progress] . >>>>>(b) >>>> >>>> >>>> >>>>OK.. Computers don't do any of that... >>> >>>Nothing of pre-recorded stuff? Mhmm. Or do you use a different wording? >> >>You want to stick to the rulebook, fine, remove all recorded chessknowledge by >>the human. It may be stored in the neurons in the brain and not on paper, but it >>says "or otherwise recorded matter", so get rid of it :) >> >>If you remove all pre-recorded knowledge in a program, you remove also the >>evaluation terms, piece square tables, all the search/extension/pruning rules >>that are based on chessknowledge, in fact even the chess rules must be deleted. >>You have no program if you choose to interpret the FIDE rules like that. >> >>-S. > >You might be caught in black/white perceptions. Why this couldn't be discussed >and regulated? Otherwise we end arguing that the progrma and the machine were >made by man. > >Rolf Tueschen I'm the one caught in black and white perceptions? Isn't it you who want the computer to follow the rules applying to humans? My point was that it isn't that simple, and if you throw out pre-recorded knowledge you have no program. The whole idea of a chessprogram is to put as much good information into it to make it the best possible player. If you say computers can't play with books, then you are giving them a disadvantage relative to humans who can memorize long opening lines. If you give them a book they are the ones with the advantage because of their much larger memory. Two remarks: Obviously it is impossible to set equal rules for two such different opponents. Fairness is irrelevant and impossible to achieve. Setting rules that specify what the computers can or can't do is road I don't want to follow, one rule will lead to the next and none of the rules will seem logical in the end. If it can't play with book, then I say: why should the computer have to re-calculate the same information again and again and again on the opening position at every single game? Humans are allowed to learn, why then shouldn't computers have the same rights? The technical way of learning for a computer is to put moves into a file, in principle the same way as humans does when they absorb opening theory. You can't use the argument that the computer haven't proved it could find the lines for itself, for that requirement doesn't apply to humans. It seems you want to punish the computer for not being human or not having specific human trades, this is very bad logic IMHO. You have to realize that the computer doesn't understand anything, it knows only what we tell it, so why should we not be allowed to tell it everything we know about chess if the object is to build the best possible chessplaying program? -S.
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.