Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: About random numbers and hashing

Author: Sune Fischer

Date: 15:50:45 12/08/01

Go up one level in this thread


On December 08, 2001 at 18:28:58, Ralf Elvsén wrote:

>Yes, since your first(?) post about the non-invarianve of the Hamming-
>distance under rotaions I have started to suspect that for some reason
>the "Hamming-generated" numbers were better in this sense (i.e. better
>distributed). But it makes no sense why it should be so...
>
>Ralf

Think of it this way, if you have three vectors a, b and c with hamming
distances resp. 1, 2 and 4 then this in itself will guarentee they are linear
independent and therefore cannot collide. If you optimize for Hamming distance
you may get distances 6, 6 and 6 (for example) and then they no longer need to
be linear independent and they might collide. It would seem as though a good
spread over Hamming distances will also give a high span of the space for each
subset of vectors.

But then again since this Hamming trick is so widely used, I wouldn't be
surprized if the litterature has a proof somewhere that it increases the
spanning (like David says).
It is hard to prove either way I think, but if I had a hamming optimizer
algorithm I wouldn't mind doing a little collision test myself ;)

-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.