Sounds mostly like cheap marketing. Change it a bit and Red Had could post it on their website and tell us that JEE 6 is better than Spring.

I think that maybe you are fearing a bit the competition. Before JEE 6, Spring was clearly better in many areas but now with CDI (which is IMHO better then DI in Spring 3) and the other improvements, it will be harder to sell Spring to developers and customers. That does not mean that Spring is bad but it is not better anymore than JEE 6.

So I think both the Spring and the JEE 6 people should talk more about programming than trying to convince others why their solution is better. Of course, with the huge amount of money involved, this won't happen.

I think one should use either JEE_ 6 OR Spring (or JEE with specific Spring solutions like Spring Batch of tasks outside the server). No point in mixing this in any way. And GWT or Wicket can also be used with CDI (or it is currently in the making, check the jboss.org website).

And Spring is not the only thing that is "evolving". There are projects like Seam 3 that provide many additions to JEE 6 (although Seam 3 is not yet final).
By the way, over 40% of the people in the US don't "believe" in Evolution at all so neither Spring nor JEE can evolve

I agree that it takes quite some time until all the stuff is final. At the moment, we cannot use JEE 6 as we use JBoss 5 and JBoss 6 is not yet final. It really takes a long time.

Also, I think that moving from Seam 2 to CDI/Seam 3 will take some effort but we really like all the new features in the standard and will probably go for it once JBoss AS 6 is here.
That said we currently also test the JEE 6 versions of our software in Glassfish 3 and have absolutely no problems.

We also use Spring Batch on the side and love it. This is very cool stuff!

I think it will be a matter of taste in the future. Either Spring 3.x or JEE 6 Seam 3 (or standalone when sufficient).
Using JEE 6 with some extra stuff from Spring like Spring Batch is currently our way to go and we like the results so far.

A few things are annoying with JEE 6, for example that some annotations are there twice in different packages (e.g. @Managedbeans). This is confusing and @Managedbean shouldn't be used anyway (use @Named). They should have cleaned up this mess before releasing the standard.

Another problem is, that some of the new JEE 6 books do not cover CDI. There is a huge lack of knowledge among developers about CDI which may be it's worst problem. I've talked to many people who like JEE 6 (Servelt 3, EJB 3.1, etc) but don't know much (or anything at all) about CDI.
This is not a technical problem of CDI (I think Gavin and the others created something really great with CDI) but slow down adoption, at least until more people know about and understand CDI.

In the end I think that if a programmer cannot create something great with Spring 3, JEE 6 won't be of much help either.

I hope my first post was not too negative. I've used Spring very successfully and never had any problems with it. Keep up the great work.

Claus

EE7 should (or rather MUST, else it may never see the light be modular!
Whether based on OSGi, Jigsaw or a bit of both, that's to be decided (and at least indirectly via OSGi or Eclipse SpringSource should also be present

So there will be less a question of "JavaEE vs. Spring" or "Seam vs. Spring", than "What is right for me?".

Many people at a recent event that still has "Java" in its name asked why and where you'd need Spring if you got Glassfish. IOC is a good example where alternatives have driven the standardization process while SpringSource and others were just watching. Tapestry is still even further behind btw, so don't worry.

A real integration of previously separate and often developped without caring about other parts of the ecosystem is what EE7 shall also bring. Then you can "Mix & Match" to some extent although many vendors currently try to lock developers into their own portfolio.

Werner
(JCP SE/EE EC Member)