I am only about a quarter of the way through the book, but thus far I would have to say the chapters on BerkeleyDB and Eclipse were the most interesting, both containing a lot of history and insight into how the projects changed over time. The chapters on Audacity, Bash, and CMake were all pretty good, but I did not find them as enlightening.

I guess that means, thus far, the book is about a third of each: very interesting, modestly interesting, and mediocre. Not too bad considering each chapter is written by different individuals.

Well at 1/3, 1/3 and 1/3, at least the selection of authors chosen are beating Sturgeon's Law.

The chapter on LLVM is excellent as well. Really explains the architecture and the choices made.