Aspi Havewala said
This book sways between discussion and dissection - addressing entrepreneurs, curious technologists and architects. It's inconsistent in its treatment of the material it chooses to add to its scope. Fortunately, this doesn't mean the book isn't useful to read or fairly enjoyable in parts.
Early on the authors take on the considerable task of explaining Web 2.0. They use an approach in which 1.0 applications are compared to their 2.0 successors or competitors. This approach does not work for several reasons.
The 1.0 application on display has evolved since Tim O'Reilly picked the contrasting applications a while ago. So there is no good benchmark to use. The authors point this out in several places, which makes the whole comparison more kludgey. Instead of an architectural analysis that is crisp, the comparisons devolve into feature discussions. So what defines Web 2.0? A set of features, the approach, its execution? The resulting discussion doesn't really come out and clarify Web 2.0 any more or less than available material on the Internet.
However some central themes to emerge in this discussion that are put to good use later.
The second half of the book presents a reference architecture for Web 2.0. Its a decent chapter, but not comprehensive. I was very unclear about how to utilize the information that was presented in Chapter 5.
Chapter 7, which talks about Web 2.0 patterns at a deeper level, is easily the most enjoyable chapter of the book. Like it predecessors, its not comprehensive, but it covers important ground.
Each Web 2.0 pattern is explained very well. There is a paragraph on the context in which the pattern should be utilized. There is material on the pattern's static structure and dynamic behavior and notes on implementation (these are a little thin for some patterns). A nice section on gotchas (called consequences) is also available.
Some patterns are a little broad - for example, is Rich User Experience really a pattern or an expectation in the Web 2.0 context that consists of serveral, constantly evolving, well-understood patterns that encompass graphic design, usability and dynamic web programming? Fortunately, regardless of the approach, the resulting discussions in this chapter are all good ones.
Bob Reselman said
The term "software system architecture" means many things to many people. Coders look at the microcosm. Systems engineers look at the macrocosm. By nature, writing about the software system architecture is hard. Most times no one size is going to fit all.
Thus, I am always apprehensive about a book that purports to explain software system architecture in general. Usually such books are written by highfalutin Big Six consultants for the benefit of entrenched IT managers that need to be "in the know". In most cases neither party has the aptitude or real world technical background to grasp the technical topic at hand in a manner that is useful and accurate. As a result, because neither writer nor readers are real world, the book flops.
So here's the good news. O'Reilly's Web 2.0 Architectures does a more than competent job of explaining Web 2.0 enterprise system architecture. The book takes a real world approach. The first two chapters are a primer in the basics of the Internet and Web 2.0 based software systems. In the primer sections you'll learn about the fundamental technology static and various architectural patterns that make up web based enterprise systems: Service Oriented Architecture (SOA), Software as a Service (SaaS), Participation-Collaboration, Mash up, Semantic Web Grounding, to name a few.
Then in Chapter 3 the authors frame real world, web based systems such as AdSense, Flickr, BitTorrent,etc..., against architectural concepts, comparing older systems against the newer Web 2.0 counterparts, DoubleClick vs. AdSense, for example.
The writers consider the ability to think and conceptualize in terms of structured abstraction as the primary tool in the architect's toolkit. Thus, Chapter 4, 5 and 6 get you thinking in terms of Models and Patterns.
Chapter 7 serves as Web 2.0 reference/lexicon and Chapter 8 provides the call to action for creating Web 2.0 architectures that last.
After reading this book will you be able to get a job at Amazon designing its next generation Movies on Demand web service? The answer is no. It takes years of experience and some very expensive failures to understand the intricacies of designing and implementing a system of such enormity.
Will you have a formalized understanding of the thinking, patterns and implementations that surround modern Internet based software systems? Yes, by all means.
In terms of the nuts and bolts of the book, there are pros and cons. The book is short. This is good. You can read it in about 4 hours. The conventions used are clear and the writing style is engaging. There is ample reference to online links. Thus, for ease of use, you'll do well to read the book in the online Safari version.
One of the drawbacks of the book is that at times, the book introduces some pretty hard core technologies and then lends fleeting reference once introduced. For example, the book devotes a single paragraph to introducing the concept of the Enterprise Service Bus. Those of us that have worked with the ESB know that it is an ornery technology that comes with a considerable learning curve, and more considerable implementation curve. Yet the book glosses over it. However, such hit and run tactics are limited.
Overall, I found this book useful. If nothing else, it wrapped formal language around concepts that I have been carrying in my head for years. And, once the reseller discounts kick in, the price is right!
I definitely recommend this book as a common reference for any team of developers making modern enterprise level software.
Comments