Author: Robert Hyatt
Date: 21:05:13 06/06/98
Go up one level in this thread
On June 06, 1998 at 23:44:20, Steven Juchnowski wrote: >Has computer chess programming reached a stage were it is becoming >progressively harder to improve performance with each new version of >software which hits the market? > I don't believe it has reached some limit, no. Progress did go faster in the 70's and early 80's, because most chess development was done in research environments, with openness about new ideas and algorithms. That came to a screeching halt with commercial chess programming became a viable undertaking. But progress is still being made, just slowly. Much more is coming from the new hardware, because it lets us do things we could only dream about 10 years ago, without going too slow to survive. >Is it now at a level that improvements are only gained by running >programs on faster hardware? hardware makes a huge difference... It makes old programs run faster, and it lets new programs do things they couldn't do 5 years ago because they would be too slow... > >Or is computer chess programming in a far more dynamic state of flux >where new ideas are constantly being developed but have not yet been >tested for improvement? >For example what about the application of "neural learning methods" in >computer chess? I hear a that such a computer chess program is >already in existance. Is this the new direction? > >Regards whether neural computing will have any impact on computer chess or not is an interesting question. At present, no. But perhaps...
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.