This is a book that deserves 15 different sorts of review. If computers touch your life, then here is a way of entering the excitement of closing the gap between technology and life, between computers and nature. If you have any sense of fun, get the book, get on the internet, download the programs and start to play your way into exploring the boundaries of science.
That is one review. Another might be one that persuades those stuffy people who write A-level computing and information technology syllabuses to read this book and use it to design a syllabus that would give teenagers a real subject. The Computational Beauty of Nature provides a wealth of ideas that are relevant, challenging, stretching, deeply absorbing, and ideal for school treatment.
Yet another review would trigger daydreams of a world we might have been drawn into had we read this book when we were teenagers. In receptive hands, this book will inspire degree choices, life choices, and a greater appreciation of computers and nature.
The book's ponderous title belies the decade's work of love that it is. A substantial selection of core subject areas are covered in 24 chapters that require a modest level of mathematics. If you do not understand calculus, you are not going to learn it from this book, but the summaries provided are useful revision, and not so tedious that they would put off less numerate readers.
Besides, it is easy enough and quite satisfying to play with the many versatile programs: the book's website can be explored at http:///mitpress.mit.edu/0262062003.html
Every chapter is illustrated by programs, from simulating predator-prey systems to applications of genetic algorithms. The programs can be run directly and the source code is available for anyone wanting to try their hand at programming.
The programs are well documented, both in the book and on the website. Indeed, the website looks like it will become a lively centre of activity for the subject, and for pedagogic developments.
Perhaps the best review of the book would see it as popular science. The Computational Beauty of Nature makes science accessible to anyone who enjoys, say, the level of Scientific American . A temptation of popularising science is to sink into reportage. That is, you watch the lives of scientists, but do not become a scientist. Fortunately this book is crammed with ideas and programs that are both flexible and powerful.
It starts with a clear exposition of computability that leads up to a discussion of Turing and Godel's ideas on what computers can and cannot do in principle; the associated program is an interactive Lisp system that can be used to explore recursion and other concepts in a general way. You can also draw all sorts of coloured Mandelbrot sets and watch flocks of birds fly across your screen.
The book takes the view that computing and natural systems ideas are unified in a pentagram of relations. This philosophy is a bit wobbly - I do not mean this as a criticism - in the sense that it makes one want to rush in and sort it out. In fact this is its characteristic strength: the reader is invited to experiment with ideas, formulas, pictures and programs - and is left on the edge of a new personal understanding, or even on the edge of genuine new discovery.
The balance is tantalising. A useful glossary, a substantial bibliography and a good index also mean that the book will make a good travel guide for the journeys it will inspire.
Harold Thimbleby is professor of computing research, Middlesex University.
The Computational Beauty of Nature: Computer Explorations of Fractals, Chaos, Complex Systems and Adaptation
Author - Gary William Flake
ISBN - 0 262 06200 3
Publisher - MIT Press
Price - £35.95
Pages - 493