Rating: 
Summary: Don't do ATL or COM without it
Review: This is simply one book that you do not want to be without when designing and implementing a COM project. I purchased the book after I found myself in my first COM project and without any coworkers with any experience in this area. As a comment, I've also rummaged around the Internet and purchase ...books in the effort to come up to speed as quickly as I can. ATL Internals is the best source of data on ATL that I've found so far. ATL Internals not only discusses ATL but also does and excellent job discussing COM's theory of operation and how a Win32 OS supports COM. As a warning, however, ATL Internals isn't the only book that you need on a COM project because it is specialized on ATL (and on COM in general). While ATL is useful it is not the only thing that you will have to deal with on a COM project. You will probably need to work with IDL files and for that I can unequivocally recommend "Essential IDL" by Martin Gudgin. ATL Internals accomplishes several things well; I will try to describe the most valuable aspects of this book. One of the best properties of this book is that it is thorough. It covers ATL in depth. The following is a partial list of the material in ATL Internals: 1) ATL Internals covers using the wizards and explains how to proceed beyond the point where the wizards quit.
2) A though discussion of ATL facilities. For example: CComPtr (a COM smart pointer), CComBSTR (a class wrapped around the error-prone BSTR datatype) classes are covered in detail. The text conversion macros are completely covered here. In other books and on the internet I saw code examples that used them but will little explantion.
3) Bug warnings, both your (potential) bugs and bugs within ATL. The authors point out the buggy parts of ATL and they invest the necessary effort to warn out about aspects of ATL that, if naively used, create bugs. Numerous pages are dedicated to showing both how to correctly use certain hazardous parts of ATL as well as examples of buggy code. Depreciated techniques are also described for completeness.
4) Discussion and examples of various COM issues, including threading issues, connection points, the Service Control manager (SCM)and so on.
5) Detailed discussion of the numerous macros that Microsoft uses to conceal substantial code, (e.g. BEGIN_OBJECT_MAP, COM_INTERFACE_ENTRY_IMPL, etc.)
6) One of the appendixes shows a cross reference between ATL classes and header file names. This was a welcome convenience (yes I could have greped, but this is quicker) ATL Internals not only has great content but it is written in a very readable style. The typeset is easy to read and the diagrams are well done.
Rating:
Summary: The best book for understanding what's REALLY going on
Review: I first bought this book because it was the only one I could find that had a discussion of ICollectionOnSTLImpl that I could understand. I was already fairly familiar with ATL - at least how to use the wizards - but found that a lot of the documentation that was available was less than useful, especially the collections stuff. After successfully using Chapter 7 (Collections & Enumerations) for what I needed, I went back and read the whole book from front to back, and started really to understand what was happening under the covers - the whys and wherefores for all of the seemingly quirky little code chunks that I had seen peppered throughout my code, but not understood fully. This book has become an indispensible resource, that I keep close at hand. I can't really recommend it as a first book on COM or as an introduction to ATL, but this is not a criticism, because that's not its intended purpose. For in-depth coverage of the workings of ATL, though, I haven't come across a better book.
Rating:
Summary: Essential
Review: Takes the background established by Don Box's "Essential COM" and shows the low level why and how of ATL. If you want to know what DevStudio's wizards are doing for you, get this book. It will help you interpet the Wizard-generated code. The author was responsible for producing ATL, and he can write well too! This book is NOT like Myers' applied C++ books, it does not apply ATL to any problems of note, it just very lucidly explains what ATL does for you and most importantly, the caveats!
Rating:
Summary: Philip Leitch
Review: This is a truly great book. It gave me all the answers I was looking for, and I was looking for a lot of answers. The section on strings (windows strings) CComBSTR is indispensable (chapter 2). I can't understand how people could rate this book badly - I tried the rest and this was the best. The only down side, as I see it, is that the book is rather dry. Then again, if you expected anything other than a dry book you shouldn't be looking in non-fiction computer books. Humour was used but it was sarcastic and rather dry - quite frankly it should have been left out completely. The book was VERY full on, presenting vast quantities of information very quickly, with enough information to make things clear.
Rating: 
Summary: Very poor book. Reads like user's guide & offers nothing new
Review: I must confess that I've read all previous reviews & am @ a loss as to what all the high ratings are all about. I've read the book & found nothing in it worth reading. Chapter 2, dealing with ATL Smart Types, was a bit interesting but offered nothing I would not have been able to infer myself via trial & error anyway. Appendix B offers a good list but is also nothing I can't get from a class diagram. I too was interested in Connection Points in ATL but found Chapter 8 to be non-contiguous & struggled to make it through it twice so as to try & understand. @ the end, I was more confused than when I started. I have, & am also currently, coding with MFC & ATL connection points. Conceptually, you need to understand what connection points are all about & then let the framework do the dirty work anyway. That is what a framework is for. From there on, you code the business logic & be done. Yes, ATL is closer to the metal but you can still use the class wizard & visual studio to get the job done. When I think of ATL, I think small, fast & true multithreading. These topics & themes were only sparsly covered in the book. Lastly, I found the examples used in the book especially poor, lacking context, & most certainly not mainstream. I would have expected this book to be far more concise & to the point & not just present a list of disconnected topics. I would have expected this book to be authoritative. I strongly feel your time is better spent otherwise.
| 
