RE: Good Books on Chaos Theory

Dusseault ("Lisa)
Wed, 19 Aug 1998 14:49:07 -0700

I'm currently reading Kevin Kelly's "Out of Control" (previously recommended
on FoRK). It discusses emergent behaviour and advocates out-of-control
systems. It does discuss ecologies and complex software systems, though so
far it has not directly linked them.

I'm only halfway through, but I'll venture to comment anyway -- I think I'm
adding to previous FoRK commentary on the book rather than repeating it.

Kelly isn't a scientist, and doesn't talk like one. Instead he communicates
like a wide-eyed amateur: "Look at this cool idea!" He makes arguments by
analogy. He covers an extremely broad selection of subjects relating to
chaos, emergent behaviour, control systems, stability. Of course many of
these subjects are only lightly touched upon, which is what makes this a
light, readable book. If you find WiReD to be annoying (Kelly is editor of
Wired) then you'll find this book annoying: jingoistic claims for the
future of cyberspace abound.

On the other hand, Kelly is very readable and the book is very approachable
and interesting. He is not a scientist, but he covers the work of
scientists. He presents compelling ideas, some of which I've started to
discuss with my friends. I'm starting to get inspired by it because my
training (engineering) was all "control systems" and in real-life, what we
usually have in complex systems is more-or-less out-of-control. IP packets
are routed by individual routers making their own decisions, and this turns
out to be cheaper than the more-controlled phone network. We push and prod
complex systems (i.e. shutting down the stock market on days with very
high-volume trading) but we don't control the individual transactions that
make up a complex system.

Kelly points out that humans are not comfortable with giving up control --
NASA's choice of projects could be used to illustrate this tendency. Most
people I know who live in capitalist countries and dislike communist systems
still wish the government would step in and "control" pollution, job
conditions, wages levels, hiring practices, resource usage, and many other
factors. Rarely do we examine other methods of achieving the same goals
besides directly controlling industry through laws. In other fields as
well, goals are first attempted through direct control, and whenever we give
up that control, it is only with misgivings.

Look at all the science-fiction stores about "out-of-control" robots. A
friend at lunch mentioned a story which he attributed to John Barnes --
about a complex mining system in space, with robots programmed to go out and
return with valuable ores. In the story, the robots had a certain amount of
self-determination, and decided that cannibalizing each other was the best
strategy for returning with raw material. In reality, systems are built
with some safeguards. Also, even with some danger of "out-of-control"
systems leading to bad situations, we sometimes have to consider giving up
some measure of direct control in return for efficiency, low cost or low
management requirements.

A lot of what I get out of this book is not directly from the book but from
the thoughts & discussions it is inspiring. For others interested in the
same thing, I'd recommend reading the book with a skeptical outlook.


-----Original Message-----
From: Ian Andrew Bell []
Sent: Wednesday, August 19, 1998 12:34 PM
To: FoRK List
Subject: Good Books on Chaos Theory

Making the Predictable Unpredictable?

Rather than actually SHOP at Amazon, I thought it'd be worthwhile to ask
the list if anyone has ever read any approachable Layman's books (that
of course also appeal to the %3) on Chaos theory.

The more I think about The Interconnectedness of Things(tm) the more I
realize that networks are ecologies and therefore are subject to
"emergent interactions" similar to biological organisms, like Joe's F1
cars. Chaos theory has wonderful applications here, no?

Ian Andrew Bell
BC TEL Interactive (604) 482-5708

"Make it idiot proof and someone will make a better idiot."