« Making Applications With Degrafa and ObjectHandles | Main | The Time Traveler's Wife »

Critical Mass

Philip Ball's "Critical Mass" is an exploration of how one thing leads to another. I found the book to be a laborious read and overall wouldn't recommend it. While it touches on interesting topics such as phase transitions, power laws, self-organizing patterns, collective motions, and scale-free networks, the book isn't cohesive. It is bookended with the thought of using the laws of nature to guide and predict human nature. While there maybe corollary between the two, I didn't find any of the law applications that persuasive.

The use of modeling, particularly the chapter on traffic has merit, but again didn't feel like it derived from natural laws, it was more simplistic rules producing complex emergent behavior. Other models explored, such as the axis versus allies assignment, felt like the model was tweaked to get the desired result bringing into question the validity of the entire exercise. Some models, like predicting market crashes, clearly have an observer effect as noted on page 236.

When the book is talking strictly about natural phenomena (meta-stable states, one way flows), and not trying to relate it to human behavior and affairs, it makes for okay reading. The problem is that between the good scientific bits, the author's diversions into politics, policy, and people feel preachy and self-righteous. Given the length of the book I'm sure everyone is bound to find some nugget of usefulness in it, but slogging through the rest to find that bit doesn't seem worth it.

Tags: book science