>> Maddeningly, I've searched and searched the online version, but cannot 
>> find any reference to this experiment. It definitely appears in the 
>> paper copy I have though.
> 
> Cool. I'll dig thru the index.

Chapter 8, "Implications for everyday systems", section title "Fluid 
Flow", page 376 in my hardback copy.

He shows a cellular automaton consisting of a hexagonal grid. Each cell 
may optionally contain a particle, which can have one of 6 possible 
velocities. (Note that all particles have unit momentum. There is no 
attempt to simulate varying particle speeds.)

When two neighboring cells both contain particles, a simple automaton 
rule defines the new velocities for those particles. (Again, casually 
glossing over a vast swathe of actual phenomina that occur with 
molecular collisions. Here there is no vague attempt at Newton's laws of 
motion, or even intuitive plausibility really.)

Viewed locally, all the particles dance around more or less at random. 
However, if you average over a very large number of particles, you what 
something looking suspiciously like fluid dynamics. Specifically, 
Wolfram demonstrates that if all the particles have random initial 
velocities, the overall net flow is zero. If a large minority of 
particles all move in one direction, and there is an obsticle in the 
way, you get trailing vorticies behind it. And if enough particles are 
made to move the same way, you get real turbulence behind the obsticle.

All this from just individual particles interacting in a very 
physically-incorrect manner. (OTOH, you have to average absurd numbers 
of particles to get this, so it's very ineffient stuff!)

>> (As an aside, I found the book to be somewhat dissapointing. But there 
>> were one or two very interesting things to be found. This was one of 
>> them!)
> 
> I thought a lot of it was pretty cool, and he did a good job of 
> organizing it into "the part you read for fun" and "the part you read if 
> you're actually going to do something technical with the results". 
> (Unlike several other tomes of this sort I've read.) Altho, yeah, the 
> hubris was rather astounding. :-)

It makes some valid points... it's just not very interesting to read.

> I trust you've read Godel-Escher-Bach?

Never heard of it.

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message