Dion Almaer wrote on usability of Java frameworks. I agree with this post. A good programming language (may be two) and a dozen of design patterns is all you need. SQL is a must. Period.
As I was younger, I was eager to learn any new software while it’s hot. Right from the oven. Now I look at things more pragmatically, and unless an XYZ-framework brings something really-really different and new to the world of if-statements, I won’t be using it, unless my client (who pays my salary) does. But I do read books/articles introducing these frameworks.
On the other hand, we’ll always need tools that are addressing specific business needs. For example, now I’m working on a financial trading application and we need speed, speed, and speed. That’s why we are considering frameworks/solutions that provide distributed cache, minimize garbage collection, optimize HTTP tunneling, support grid computing. And there are such products on the market. These products are concrete.
Generic frameworks will play their role in the Java martket, but do not expect any of them to dramatically change the way you program. The amount of code you need to write will stay the same, but if earlier it was 90% of Java and 10% of XML, it may become fifty-fifty or even worse.