This is a delightful introduction to applied chaos theory and information theory, written by a physicist for other physicists. I found it very accessible, unlike my previous encounters with information theory which were entirely impenetrable.
It's something of a classic, apparently, and I can see why. Not only is it well thought out, planned and executed, there are some delightful turns of phrase, too, e.g: "The longer a string of text the easier it is to predict the next lette."
How often can you say that about something the author repeatedly refers to as a "paper"? (It's 111p of main text plus appendices...) It is also charmingly set in some ancient mechanical type-writer font with all the equations requiring Greek or integral signs hand-written in before going to press. It's both a shame and a testament to its usefulness that my copy is beaten up and verging on falling apart...
More technical than a pop science book, but also less technical and more expository than a paper or textbook. A really neat little book that gives a quick introduction to chaos and information theory. Topics touched on include entropy and stored information as measurements of system predictability, periodicity and aperiodicity in phase space, high dimensional systems, and modeling.