Conflicting reviews

March 28, 2010 12:40 AM

Wolframm's "A New Kind of Science" is a very thick book that would require a significant investment of time to read through. I'm interested in the computational modeling of complex adaptive systems and have heard it recommended as an interesting read. I've also seen it described as a huge load of crap. Is it worth the bother?
posted by moorooka to science & nature (12 answers total)

Well, strictly speaking "Interesting Read" and "Load of Crap" are not mutually exclusive.
posted by delmoi at 1:44 AM on March 28, 2010 [6 favorites]

I didn't enjoy reading a copy borrowed from the local library, so I wouldn't call it an interesting read. Interesting idea, perhaps, but too much wading for me.
posted by anadem at 2:44 AM on March 28, 2010

I own a copy, and I found it worth reading, but I also find cellular automata interesting in their own right. I haven't read it since just after it was published, though (7 years ago?).

I mostly skimmed the parts where Wolfram goes on about how the universe must be a giant cellular automata and how cellular automata will replace old paradigms as the new way of doing science; those bits weren't all that interesting and, really, are beyond the scope of the book. Those parts are, I think, what most people complain about.

The book should perhaps have been titled "everything there is to know about cellular automata" --- that would have given Wolfram's ego something without the grandiose claims of being a broad revolution.
posted by joeycoleman at 3:20 AM on March 28, 2010

I love reading about cellular automata, so the book was interesting on its exploratory basis. I didn't get through all of it, however. If you like CA, it is one of the major stepping points.
posted by Blazecock Pileon at 3:26 AM on March 28, 2010

Also, what do you want to know? "computational modeling of complex adaptive systems" is kind of a vague phrase. Do you just want to know more about Cellular Automata? Or do you want to know more about things like Turing machines and the like? Or about machine learning? These are all interesting topics, for sure.
posted by delmoi at 4:28 AM on March 28, 2010

I'm interested in a clear explanation of the similarities and differences between coupled map lattices and cellular automata, and the relationship of either to statistical mechanics. As you might be able to guess, I don't know much about anything really, but I have to choose a topic for a research project.
posted by moorooka at 6:05 AM on March 28, 2010

I used to work in complex systems. Wolfram's book is entertaining and worth a quick skim if you want a layman's introduction to one small slice of complex adaptive systems. Ignore the headier speculation about how the universe is a giant CA. Entertaining, but without any meaningful science to back it up. The good news is the book is only half as long as it looks: you can generally skip the end notes. The worst part of the book is that it has absolutely no bibliography, no references to the actual research Wolfram is talking about. Also it ignores most productive complex systems research focusing instead on toy models he prefers.

For an insider review of the book that is harsh but insightful, check out Cosma Shalizi.

This would be the part where I'd recommend some alternate reading or syllabus for you. Unfortunately I'm some 15 years out of the field. One place to look is the courses taught by University of New Mexico professor Stephanie Forrest. She's a leading researcher in the field and a great teacher. Another place to look would be University of Michigan's Center for the Study of Complex Systems. And the Santa Fe Institute has been a major center of complex systems research for nearly 20 years, but since it's not a teaching center you will have a hard time finding introductory material there.
posted by Nelson at 7:40 AM on March 28, 2010 [1 favorite]

Ooops, just read your clarification. I wouldn't look to Wolfram's book to help you with such specific questions. You'll have to dig a level deeper to find specific papers written about coupled map lattices, etc. There's a wealth of very solid statistical mechanics work out there, as well as lots of good mathematical formalism on various computational models like CAs, etc.
posted by Nelson at 7:42 AM on March 28, 2010

I would recommend looking at Dan Rothman's work as he has a book entitled "Lattice-Gas Cellular Automata: Simple models of complex hydrodynamics".
posted by a womble is an active kind of sloth at 7:51 AM on March 28, 2010

I would recommend it. As to the length, approx 1/3 is footnotes, and there are a ton of graphics. Despite its appearance length is not a major concern in reading it.
posted by norabarnacl3 at 9:24 AM on March 28, 2010

For such a lofty title, the only new science presented in the book is a proof that Rule 110 cellular automaton is Turing-complete. Matthew Cook developed the proof.

I can't help but ask how has the book revolutionized science in the 8 years since its publication?
posted by malp at 9:43 AM on March 28, 2010

Nthing delmoi: It is an interesting read that is a load of crap.
posted by DU at 11:50 AM on March 28, 2010 [1 favorite]