Customer Reviews

COM+ and the Battle for the Middle Tier

5 star:		(7)
4 star:		(4)
3 star:		(1)
2 star:		(2)
1 star:		(3)

 

 

It is unclear weather the book is a marketing or technical document. For example, the comparison between COM and other frameworks is completely biased. The book doesn't deal with the following issues: 1. Object-Relational automated mapping tools 2. The problematic nature of a component specification dictated by a single vendor 3. The problematic nature of committing to COM, especially due to Microsoft's anti trust legal issue and the strong ties between COM and NT technology.

I liked the Don Box COM books Effective Com and Essential COM much better.

Overall rating: OK

When I first started reading this, I thought it was a comedy. Imagine comparing eCommerce to Starbucks operations! But, then I figured out, Roger was so confident in his grasp of all the technical concepts, that he did not need to be worried about how non-technical it appeared to his readers. You don't write the way he does unless you have great depth of understanding of the issues. He gives psuedo-code examples of every issue. His analysis of CORBA, Enterprise JavaBeans, and COM+ (MTS) is fair and insightful, allowing the reader to see the similarities as well as the differences. This is indispensable if you want to make a choice now about how to program web stuff. You don't want to start out making database connections from your client just to get the boss off your back, if that approach won't handle any more than a few concurrent users, no matter how quickly you can get the site up. You need to know why with some COM platforms you don't need to worry about threads, and pooling. Unfortunately, if you listen to the vendors of the platforms, you won't find this information, until too late. Read the whole book before you decide if you're sceptical, and then you'll see how practical it is to get your thinking straight before you start coding.

This book provides the component that is missing from most programming books: the WHY. WHY should COM+ be used, WHY is it better in this type of application and not in that. Too many programmers buy "how-to" books and begin applying a technology before they even ask these questions. Mr. Sessions has done an excellent job of answering these questions, but don't buy this book expecting it to explain the mechanincs of applying COM+. This isn't its purpose.

Programmers should know why they are using a technology before they apply it. This book will help programmers and managers understand how COM+ is positioned against the other middle-ware technologies.

I disagree with the person who only read the first chapter. In it he describes the difficulties in developing Enterprise Software Applications. He goes on to explain in later chapters how Comware addresses those issues. I found the level of detail to be perfect. Roger starts each chapter with the history of the technology, what real world problems it is trying to address, and then the specific technical details. After reading this book you will be able to answer questions in laymans terms what Com, Dcom, Com+, 3 tier architecture etc, is.

For years, I have been trying to figure out what all the OCX / ActiveX / COM / COM+ / DCOM / dot NET / whatever-they'll-call-it-tomorrow stuff is about. I've succeeded in developing ActiveX components and also software to use them, and in so doing put bread on my family's table.

But I've never really had a top-level overview. To use old-time IBM jargon, I've never come across the book called "Theory of Operation" for the Microsoft component stuff. So, to me it's always been a blizzard of complexity. It mostly works, but mere mortals like me don't get to know why.

Yesterday I stumbled across this book in a public library, and sat down to look at it. I ended up reading most of it.

It is a good book. My criteria for good are:

--readable -- yes, very.

--not written to be propaganda. Many books about m'soft stuff start out with "this is the most important advance in computer science since von Neumann" or some such breathless rubbish. This book starts out by saying the author was skeptical about his subject matter but found real value.

--opinionated. The author IS opinionated. But he discloses his biases and justifies them. That's so much easier to read and absorb than gray and neutral writing. It DOESN'T mean the reader has to agree with the author.

--demystifying. The author cuts through the bafflement that M'soft's marketing department has created by their lousy choices of rapidly changing terminology, and shows the strengths of the m'soft component model. I have been unable to cut through that bafflement on my own, so I'm grateful.

Worth reading!

You wouldn't think of it by reading the title, but this book is actually quite crazy. I suppose that is the best way to describe it. I've never read a computer book like it. It's both technical (on a system architecture level) and free form (stream of consciousness). There are enough jokes here to make you laugh out loud, and enough bad jokes to make you laugh just the same. And, you get a unique viewpoint on all things COM. Roger Sessions says some on-the-money one-liners throughout the book, but you'll notice alot to disagree with. For example I found his remark that "no one has ever figured out how to make use of inheritance in OOP" way out to lunch.

I recommend this book as a refresher companion to all those dreary COM manuals out there. It's also a great political primer. You just know some Javabean fanatic is gonna annoy you someday with his rantings. You'll know how to tell him off after you read this :)

The only thing I got out of this book is that Mr. Sessions drinks a lot of coffee at Starbucks. I think the purpose of this book was to promote Starbucks to programmers and maybe this is how Mr. Sessions will make his money. A word of advise to Mr. Sessions. When addressing programmers and technical manager you *can* use real world examples. You are not explaining things to children here. I was very offended by the shallow examples the book gives. Additionally, it will actually help to write a book on COM+ if you have developed COM+ applications. Reading the help files do not qualify you to be an authority and write a book. This book is a waste. Luckily, I was able to get my money back.

It is clear that Sessions is biased toward COM+. I would have no problem with that if he were up front about it. The book provides some good COM+ insights and recommendations, but it is like going to Starbucks and asking them their favorite coffee shop.

I loved the opening chapter on Starbucks. I also enjoy the fact that some here found that stupid and patronizing. This is the charm of the book. Reading chapters is like spending dinner in the company of a top industry veteran who's been through all the battles, switched sides along the way and can tell you not just what, or why, but how things got the way they are, in a very lucid way. It's filled with haymakers like (not an exact quote): the arrival of MTS in 1996 instantly made the COM programming model obsolete. In short, he keeps you reading. And he does know COM+.

For those with a rudimentary understanding of middle-tier architecture, this is a great book. I read it this weekend and will now be able to intelligently discuss the topic. Sessions gives some good reasons for using what he calls COMWare, which consists of COM+, CORBA, and EJB. His guidelines for effective use of COM+ will save us many hours of frustration, I am sure.

However, I would not call his assesment of COM+ unbiased as others have here. Just read Part III, "Competition" to verify this. He brings out some valid points, like why entity beans are probably not what you want. But his presentation style is definitely biased.

Overall, a very good book. I would highly recommend it to anyone considering one of the three COMWare solutions.