I bought this book when I had very little idea about embedded Linux systems. I tried reading it but found that its explanation of all available options / flavors, while informative, was unhelpful in trying to answer the question "what would i use to make a system of my own?". E.g. it talked about cramfs, jffs2 etc. but didn't quite address the issue of why i would choose one over the other and under which circumstances: e.g. a typical development system should use rootfs over NFS to allow for rapid iterative development and then switch to a flash based / sd-card based system for deployment (cramfs / jffs2 depending on the space constraint). A similar argument extends to uClibC vs. glibc etc.
Over the years, as I've gained experience with several embedded systems, the book's collection of all terms makes more sense, but more like an encyclopedia and a reference. I feel it still doesn't provide enough guidance on what would make a good embedded system: if i selected from the options presented, say cramfs on MIPS booting off sd-card, would i be tying locking myself into a hole? this information is better gained the hard way: looking at what platforms are already available and how active the support groups are for these.
Also, in the recent years, OpenEmbedded (OE) seems to have a strong developer push behind it. This book doesn't cover it at all.
I think the information in the book would be best complimented if the author paired the book material with a system that the readers could buy and build on their own as they read through the chapters. Yes, it would only be one specific selection from all options the book talks about, but I believe the process would be much more enlightening.