Author: Dann Corbit
Date: 18:05:21 06/17/04
Go up one level in this thread
On June 17, 2004 at 19:32:17, Dieter Buerssner wrote: >If anybody wants to implement "bitbases", and really doesn't want to do it more >or less with own ideas, the articles of Ernst Heinz show really everything >needed. After that, it should only be a programming exercize. The articles are >avialable at http://supertech.lcs.mit.edu/~heinz/node17.html > >Especially interesting in this context: > >Space-efficient indexing of chess endgame tables. >Knowledgeable encoding and querying of endgame databases. >Endgame databases and efficient index schemes. >Efficient interior-node recognition. Yes, I have those papers. >But with knowledge of the basic ideas, who to use symmetry to index >TB-positions, many people will be able to come up with similar ideas, that >should also work well (and perhaps identically). It might be more fun. I sometimes wonder if there is a fundamentally better way to compress. I have thought about a minimal perfect hash for every legal move on a sparse board. We could encode it in a minimum number of bits, amounting to the number of possible total moves for a set (as reduced by reflections, rotations, etc). Of course, you would have to compute a new hash for each new set. The reason I brought it up was that I think it would be nice if there were a standard way to do it. Imagine if every chess program author made his own tablebase files. Sounds like a clever thing that would encourage innovation. That's on the one hand. On the other hand, there would be 250+ redundant 7 gig file sets on my computer. That would be a bit annoying.
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.