Expert One-on-One J2EE Development without EJB [ペーパーバック]

I've read this book several times since the day it shipped and I have to say that this is an excellent book for anyone working as a developer or architect working in the Enterprise Java arena. I absolutely love this book given my bias - I guess I should start by stating my bias. EJB bashing is a favorite past time of a lot of people. I happen to love EJB's, with the exception of entity beans and think EJB's are a great way to create software solutions are remotable, loosely coupled and powerful. I will agree that EJB's are way too complicated with all the stupid artifacts that you need to create to create and deploy an EJB. Having worked with EJB's since 1999, I guess I am so used to all of nuances of EJB's, I can write up deployment descriptors in my sleep. Having said that, I approached this book with a little apprehension as I hate these EJB-sucks book that don't really offer any intelligent discussion about the shortcomings of EJB nor do they offer a viable alternative. Another assumption I brought to the book was that this was just a Spring book with a little EJB bashing thrown in for good measure.

To my pleasant surprise, Rod Johnson and Juergen Hoeller have written an awesome book. This book does not take cheap shots - Instead there is a intelligent, thought provoking discussion about the pros and cons of EJB. In fact, the first 120 pages (Chapter 1-5) are just a great breakdown of application architecture with a through treatment of EJB. I loved this section and re-read it several times and I found myself agreeing with pretty much everything in this section. I would equate this to a great meaningful discussion you would have with someone who really understood application architecture and development and you could debate the pros and cons of the many alternative approaches that exist today.

Chapter 6 starts the discussion of Lightweight Containers and the idea of Inversion of Control (IoC). This is not a chapter on Spring; rather it is an overview of Inversion of Control and strategies like Dependency Injection in the context of Spring and PicoContainer.

The next chapter offers a quick introduction to the Spring Framework. As everyone already knows, the Spring Framework is a very popular open source application framework created by Rod Johnson. The co-author Juergen Hoeller is another lead developer of Spring. The chapter is Spring is fairly light and people hoping for a in-depth Spring tutorial will be disappointed. Instead this chapter offers a rather high-level overview that will get you some basic understanding of the Spring Framework. I guess it's hard to cover Spring in 43 pages.

After the cursory introduction to Spring, the book moves into Aspect-Orientated programming (AOP) concepts. This section starts with a very introduction to AOP before jumping into AOP implementation strategies. After a brief discussion of AspectJ, AspectWerkz, and JBoss AOP, the authors move into SpringAOP. After AOP, the books moves into Transaction Management where current J2EE approaches are discussed and then contrasted with the Spring approach.

Review trimmed to comply with Amazon's review guildlines for length. For more details, check my blog at j2eegeek dot com.

Rod Johnson is doing a great service to the J2EE technical community with his books. This latest book is definitely a myth buster, that I was personally looking for.

I will tell you right away that this is not an anti-EJB book
that tries to prove you a case against EJBs. This is not a
cheap "Spring" framework promotion book either. This is a very mature expert one-on-one advice that is well worth getting.

Rod gives you a nicely rounded manual how to architect solid J2EE application using the latest and greatest practical solutions available both through the open source and JSR community. He propagates two extremly important ideas:
Lightweight containers and (simplified) Aspect Oriented Programming. Moreover, ha makes a very strong case for the application of Inversion of Control principle (IoC) in your applications. If you are not familiar with IoC: I see it pretty much as a savior to a J2EE technology. J2EE grew incredibly big, complex and fluffy in the recent years, and is at risk of being outflanked by more simplistic .NET solutions.
IoC offers "back to basics" approach where you as a good OO architect focus on the solid business domain model without poluting it with the infrastructure code. Through IoC supporting methods (such as Aspects) you then externalize the infrastructural pieces (transactions, pooling, persitence, logging, auditing,...) that make you apps run in the enterprise environement.

Rod's book gave me a very good basis for the creation of my own state-of-the-art J2EE solution and I am grateful for it. It is the best thirty-some dollars that I spent in the long time.

One more thing, this book in NOT a re-write of his previos book "J2EE Design and Development". I have both and they are not the same. I think you have to have both on your bookshelf in order to get the full treatment.

This book builds a great case for an EJB-less architecture, the arguments and points are layed out very carefully and very well. A different architecture is presented, with plenty of help on how to get the benefits of an EJB container without the pain. Unlike many other books in which the author seems to hope you will take his or her advice simply because it is in print, Johnson and Hoeller back everything up. The book flows well, and contains very few mistakes that I have noticed (and those were very minor editing slips). I hope to see more from these authors. In the meantime, I guess I can chuck all those EJB patterns books on the shelf and just put this in there...