Essential COM [Paperback]

I've re-read it twice since then. This is not because I didnt understand the book the first time -- its because there are so many layers of information in the book.

Some authors seem to be able to write 20 books on a subject -- You've seen the authors that do 3 C++ books a year, or 3 COM books a year. This is great for publishers, and the successful "serial author". (I am aware that Stroustrup has 2 other C++ books and Don has taken part in Effective COM -- but they're not on the same subject material)

Don takes a different approach. He's only going to write one COM book and do it properly. He does. This book is the most dense in terms of giving solid information to help you understand whats going on. Whats going on when your product is about to ship and there's only 3 "showstopper" bugs left. Thats when understanding it properly matters. It also matters when you want to design something. These details also matter to VB projects when they ship, or when they're being designed.

I generally read any COM book I see sitting around. So I've probably read about 6000 pages on COM. I've seen about 1000 wizard screenshots. I've seen 40 analogy-riddled COM explanations. I've seen "cute and funny" examples. This is the only one I've bought and I've never regretted it or covetted my neighbour's book.

Similarly with C++ books, there are authors who sell a rehash of the same material 20 times. I read any of these lying about too -- another 10,000 pages with 50 useful per book. I only _bought_ Stroustrup.

It's not (nor was it intended to be) _the_ tutorial. It doesnt have screenshots. Its not 'funny'. Depending on your initial level of knowledge, you should probably read one or two other books such as Inside COM (or intros to other books of that ilk) -- in fact the best primers are probably articles in MSDN. Use these as a primer as necessary. When you've read 1500 pages of these, come back to Essential COM, and you'll have any gaps in your knowledge filled in.

If I was allowed to own only 3 books on C++/COM development, they would be Stroustrup, Box and Effective COM. (Although Mr Bunny's Guides would come close :)

Finally, would people, regardless of positive or negative opinions held, please be kind enough to give their identity -- IMHO it invalidates your opinion if you're not willing to stand behind it.

There are dozens of COM books of the first type. There's only one of the second. This is it.

There are a lot of books that will tell you how to develop COM applications. Most of them are aimed at the journeyman programmer who is capable of following directions and making small extensions to cookie-cutter sample apps. They serve this need quite well. Don's book isn't aimed at that sort of person, and doesn't fill that need.

If you need to understand why COM looks the way it does, or how COM grew, or what its underlying philosophies are, or if you need to deeply understand the whole COM paradigm, Essential COM contains the answers.

COM is full of jargon; terms that convey only a small amount of information themselves. The richness of an object model relies upon the abstractions that serve as its foundation; if you don't understand the abstractions, you can't really apply the model creatively.

Sure, the sample code gets pretty deep, and does so rather quickly. Take the time read the samples, and understand them; consider them "exercises for the reader". There is much to be learned from them.

Yeah, the macros get deep, too. COM is layer upon layer of abstraction; the implementation is layer upon layer of code. Macros help flatten that out; they keep the programming level of abstraction high enough that one can see a forest rather than an endless collection of trees. Understand them when they're introduced; take the time to work through the code. Then use them without worry.

Is this book truly "essential"? Depends on what you think that means. Does it describe the essence of COM? You bet. Can one write COM apps without it? Probably. But they won't be the most efficient, or the best designed, and they won't make the best use of the technology.