Today I went to see the session about J2EE 6. And I saw what I expected to see (not that I read anything about it upfront): an even more lighter API, very much based on annotations, the JPA 2 spec which still doesn’t give me all the cool stuff I really need when it gets rough (e.g. replacing/extending legacy stuff on existing databases) although the Criteria API is quite nice. A nice surprise was the validation framework, also annotation based and maybe one of the most important things: a lightweight standalone EJB container that can be started from within unit tests (a bit like you do with Spring).
But, I wasn’t that happy with all this new stuff. Most of these things can be done these days already with existing frameworks, in some cases even better than what they came up with now. So again, nothing really innovating is added to the spec. Secondly, trying to avoid all the XML clutter, which I hate too (but that’s more because of a lack of tooling support imho), all configuration is now done with annotations. OK, it gives you compiler checks and it is refactor safe, but is it still readable when 5 or 6 annotations are places above one simple property? I’m afraid not.
In the afternoon, I went to see full-text search for Hibernate. Emmanuel Bernard did a pretty good job explaining what is all about and how it should be implemented. The goal of the project is to make an API layer above the Lucene engine that works just the same as the Hibernate Query API. If I ever need to do full text search, I’ll certainly give it a try.
Up to day 3, the first conference day..