Author: Robert Hyatt
Date: 06:03:05 05/11/00
Go up one level in this thread
On May 11, 2000 at 08:44:51, Hans Gerber wrote: >On May 10, 2000 at 23:15:56, Robert Hyatt wrote: > > >>First, I was _the_ author, so no, there would be _no_ code I didn't specifically >>either write, or validate, myself. Harry and Bert were both long-term friends >>that I knew very well... >> >>There was no outside influence there, most likely. IE it would be possible for >>a Cray engineer to do something as we were running on one of their computers at >>the corporate computer center. But none of the three of us could do anything >>since we were all present... unless we wanted to cheat of course. >> >>I had a strong suspicion, yes. A suspicion that I had a serious bug in the >>parallel search. Never a thought that someone else had exerted outside >>influence in any way. >> >>But the point? If I had been accused of cheating, how would I have _ever_ >>disproved the claim? I couldn't reproduce a key move. Of course, my opponent >>couldn't have proved that we cheated either... so there you go... back to >>square one... >> >>Programs have bugs... they have non-deterministic behavior... they are operated >>by humans that might or might not be dishonest... they can be influenced by >>people outside of the operator.. outside of the game room.. outside of the >>country even... I don't see how the random variables can be eliminated. >> >> > > >I have to keep down my strong emotions so that my thought process is not too >much disturbed. > >I think that your report should be analysed deeper. > >The point is that you as a scientist can handle such a situation of twilight and >almost paranoia but let us think of much weaker and less educated individuals. >Let me not write down all the consequences. > >At the same time I'm thinking about, say, Fischer in 1972. The Russians >searching for a fly... > >Is all that possible confusion _not_ sufficient for scientists to make extreme >attempts to find solutions of control? If _you_ as a programmer already invest >your heart blood into your creation how about the situation of a genius trying >to play his best chess? The situation is absolutely paranoid if any idiot could >gain control somewhere in the system. Likewise the computers and the environment >of a human chessplayer. I would like to be able to control this. I would also like to be able to control the weather. Either is equally likely. How do you prove that a program has no more bugs, before playing it? How do you eliminate non- determinism when it is caused by hashing, by parallel search, and by hidden bugs? > >To know that in principle each act could be counter-acted doesn't mean that we >can't do something to protect us. Trust is not enough. At least not when it >comes to the world championship ...
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.