Author: Mogens Larsen
Date: 08:08:42 04/26/01
Go up one level in this thread
On April 26, 2001 at 10:17:28, Jeroen van Dorp wrote: >I don't believe it. Basically you are stating that -like some people suggest >happens in particle physics- suddenly ideas change in the same direction all >over the world amongst computer programmers. > >There's no need to resort to these kind of off-limit comparisons. >It is a logical development, no new or shifting paradigms. > >We all know what good chess consists of, yet we don't know how to translate it >into a chess program. It is obviously that every programmer tries to correct >factors like "event horizon", "lack of positional insight", "being unable to >make a plan". > >Until now the bypass is "fast counting" or "a lot of chess knowledge" into the >program. But basically the principle is still the same: look for the best bypass >there is, as long as we don't know exaclty how we should translate human insight >into a computer chess program. > >Programs like CS Tal II, Gambit Tiger, Shredder 5, Deep Junior are evolutions of >the initial bypass, they don't use new or shifted paradigms. > >They still count. They still see positions, and give them a value. They still >have their horizon. > >Faster hardware enables more knowledge against a higher node speed, as an >example. The program seems to understand more, but still doesn't understand. It >counts faster, can handle more instructions per second, focuses maybe on other >chess knowledge rules. > >A new paradigm is not "a better bypass". > >These programs are getting more amazing every new day, but still it's making >things better step by step, with the same basic principles. That is true. The "translation" attempt IMO is a continuous development of a model with the goal of solving or compensating for the percieved disadvantages inherent to a computer compared with the human perception of chess, while exploiting inherent advantages as much as possible. By solving the disadvantages you'll likely end up weakening the strengths. The neverending knowledge vs. speed debate. I don't think there'll be a winner of this "contest". More likely they'll meet somewhere in the middle, because "dumb" programs tries to get smarter at the cost of speed, while "smart" programs tries to get faster or at least not slower. The general framework of a computer chess program, at least as far as I know, are very similar. This doesn't leave room for the hypothesis of discontinuous development, paradigm shift, evolutionary jumps or whatever you choose to call it. I guess you could say that the hardware functions as an evolutionary crane to facilitate knowledge jumps. Problem is, that the discussion existed before the hardware gain and it will benefit both. Regards, Mogens
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.