Now, upon very close reading, I can see how so much contradictory information got into this book: There were two authors who worked on different chapters and/or sections, but they did not work closely with each other to keep from contradicting each other. The editors did a poor job in recognizing contraditions. And it seems that no one worked on providing continuity in the textbook. In the Preface, the authors thank their reviewers for their assistance, but they openly admit that this was done "often under short deadlines".
The worst mistake in the entire textbook is Chapter Eight (page 237), where they state: "Second, unlike bridges, switches don't learn addresses; switches need to have addresses defined explicitly." (They are speaking of switches that are used in LANs.) As we should well know, this kind of switch learns its address table, just as a bridge does. Could it be that the authors have confused a regular (Layer 2) switch with a Layer 3 switch?
Here is a key concept that it not presented well: There is poor coverage in this textbook about how data packets are sent from one protocol layer to the next by adding/removing headers and trailers to the data packet, and leaving what was given to them otherwise unchanged.
There are frequently concepts that are used without being clearly defined or explained. For example, it is very difficult to find out about packets or packet-switching. Shouldn't this important topic be explained in Chapter 1 or Chapter 2?
Anyway, I do not recommend this textbook at all, and I have selected a better one for the next semester. Interesting thing is that this is a textbook that was poor in the second edition, but the new third edition is an Order of Magnitude improvement over the previous one.
I just finished reading with some dismay the avalanche of negative complaints about the book. I haven't read the 7th edition, but I read the 6th edition and have a sense about the book (and by the way, note that amazon many comments are directed to that edition and not the most recent), and I have a pretty good handle on what the book is like. It looks like a school textbook because that's what it is. It's sort of boring-looking, and it even has exercises and case studies. When reading it, one is tempted to think that information is watered down, or the book is out of date or that it is useless. Because the book doesn't give a step-by-step account of how to do things, one is tempted to think it's not a good book.
But initial impressions are deceiving. When I read the 6th edition in 1999, I initially felt precisely the same way about the value of this book. (I suspect most of the negative comments here came from people who had it forced on them by a teacher for a class). The book does NOT teach you how to network or how to configure TCP-IP or how to set up your home network. What it does instead is treat pretty thoroughly communications technologies, protocols, packet technology, switches, error-checking, backbone. It discusses everything very simply and in very clear language. Also, it covers the diversity of technologies out there, including many legacy ones (and the reader may initially see this as a sign that the book is out of date--the reverse is true; this legacy information is often hard to find in a single place).
I have bought several books on linux networking, home networking, windows networking and computer security, and I can safely say that this book contains lots of stuff you don't find elsewhere.
The writing style is rather bland (I'll admit), and some of the cases studies are not that challenging. Also, it's frustrating to read because it doesn't provide any OS-specific info. That is not necessarily a disadvantage. A lot of the OS-specific books become fixated on one particular implementation, and don't do a good job talking about general things like protocols. But this book does that very thoroughly; it has great diagrams, a very logical flow and a way of making the most difficult concepts seem crystal clear. (I feel I can comment on this; I work as a technical writer and instructional system designer).
One concern I have is that the 7th edition might not contain enough information about wireless or ipv6 (in the 6th edition, they were not covered very well, but hey, it came out in 1998). I suspect the 7th edition will include significantly more information.
I'll be the first to admit that reading this book alone won't teach you how to do that home network (for that I'd recommend "networking by example" by barry press) The book doesn't give you an immediate payoff; on the other hand, I can safely say that over my two years working at dell, I've been pleasantly surprised how useful the information turned out to be (especially the stuff about error checking and data transfer.
So, to the students out there who are griping about this "useless" and overpriced book assigned to them, watch out; in two or three years you may be singing a different tune.
IN SUMMARY: You'll need another book about OS Networking, but this book will give a thorough overview on lots of things.