Re: Similarity of Structure

From: Daniel Burfoot (
Date: Wed Mar 19 2008 - 17:06:18 MDT

On Wed, Mar 19, 2008 at 1:33 PM, Lee Corbin <> wrote:

> What do you think of pushing that idea for 2D patterns on a Life Board?
> (Easy, I would suppose.) What about for two grains of salt? Think that
> it can consistently and advisably be stated also in terms of Komolgorov
> complexity? What about large 3D or 4D objects in general?

The string based similarity measure works to the extent that you believe the
object is completely described by its bit-encoding. For any object that is
"purely" informational it should work rather well. Indeed there is a growing
body of research on the subject of using compression programs to measure the
similarity of things like musical scores and genomes. Do a
scholar.googlesearch on "clustering compression" or look at Paul
Vitanyi's home page.

Of course, the main downfall of the method is approximating the Kolmogorov
complexity. Nobody really knows how to do this well. However, if great
advances are made in the future, it may become possible to make real
scientific statements along the lines of "Mozart's music is more similar to
Picasso's art than to Monet's".


This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:02 MDT