Turtles, Termites, and Traffic Jams: Explorations in Massively Parallel Microworlds (Complex Adaptive Systems) (Paperback)

The book focuses a great deal on workings of the StarLogo programming language, which is not included but is downloadable (more on this later). The programming language allows users to simulate massively parallel systems. The book includes many code samples, programming notes and descriptions of how the simulations progress at run time. Discussions of resulting organized behaviors lie almost completely within the scope of the software simulations, but are very interesting nonetheless (although it will likely leave you wanting more). After only the first simulation (regarding slime mold), I found myself at the computer to download the software. Which brings me to my next point...

You won't find the software at the location specified by the book. It appears that the original StarLogo language was written for the Mac, and was renamed MacStarLogo. (Substitute "macstarlogo" for "~starlogo" in the URL to find the software.) When you get to the Download page, you'll be able to download the software for PC and UNIX as well as for Mac. The PC version (and version 1.1 for Mac, I assume) seems to use a newer or upgraded set of commands, so you'll have some difficulty getting the code in the book to run. The new Java-based interface, though, is very cool - it allows you to place buttons, sliders and other tools to control the simulation and dynamically interact with the program in real time. Excellent for exploring these microworlds!!!

The book also discusses a lot about the author's interactions with children while developing StarLogo programs. I found these discussions very interesting, but they seemed to focus on how we like to perceive organized behaviors as centrally controlled (versus individually controlled). As a result, much of the book was about why a non-centralized perspective is important rather than how organization is actually formed from non-centralized communities.

Overall it is a very interesting and well-organized book. Only three stars because (1) it wasn't what I expected - perhaps the subtitle would have been more descriptive as "analyzing simple computer simulations where organized behavior results from systems with no centralized control," but I guess that would have been too wordy. And, (2) the software was not easy to find, and it was not fully compatible with the code in the book. (A version of the software compatible with the code in the book shoud be made available - even if it's since been upgraded.) And finally, (3) the book seems to be rushed toward the end. (The last chapter, for example, where the author "looks ahead" is only two pages long.)

Overall, it's a great book, and it inspires a lot of thinking, but it left me wanting a bit more...

"Termites..." is about how complex behaviors can arise from very simple systems, and to that end Resnick provides a number of case histories and simple programs that demonstrate how conceptually complex systems can be simulated using only a few rules. Phenomena as diverse as the movement of traffic james, pile making by termites and the migration of slime molds can all be simulated in Star LOGO with very a few assumtions. But Resnick's programs aren't just simulations; they're models of the real underlying processes that govern these complex phenomena.

Resnick hasn't just created a clever program; he's provided a wonderful tool for exploring complexity, and found a way of translating complexity into something a child can understand- while still fascinating to an adult.

The use of Logo is both a strength and a weakness of the approach. The strength is that the code is concise and easy to understand. The weakness is that there is only one source of the software, and anyone wishing to try it is limited to the available download. This would not be such a limitation if the book described the same version, but unfortunately things have moved on a lot since the book was written, and few (if any) of the examples will work without alteration.

As well as the development of the StarLogo system, the book covers experiments in emergent behaviour. Typical sections include how parameter and environment changes can affect the growth and development of simulated ant colonies, and a theoretical basis for those "phantom traffic jams" we have all experienced.

This book is certainly interesting if you are interested in developing parallel software simulations, or if you are interested in marginal computer languages, but don't expect the code to work without effort.