JSR 330 (Dependency Injection for Java) Accepted by the JCP

ALT DESCR

The JCP EE/ES EC has approved JSR 330 (Dependency Injection for Java). The vote was 14 YES, 1 ABSTAIN (Red Hat) and 1 didn't vote (Nortel). Sun, Ericsson, IBM, Red Hat and Oracle all requested coordination between JSR330 and JSR299 (WebBeans). Check out the Vote Comments for the different positions.

Better late than never, I guess. Everybody (JCP as well as JSR 330 submitters) have committed to transparency, so we will be able to follow-up the evolution of this story. And consider Joining the JCP to influence this and other standards in your platform.

Comments:

I'm glad to see there is a lot of concern about consistency between JSR 299 and 330, and also concern about the inclusion of annotations without a container inside the JavaSE environment. I sometimes wonder if this is SpringSource's tactic to make JSR 299 not be part of Java EE6. JSR 330 says they think they can go final within 5 months; in time for inclusion in Java 7. Now they have no choice but to work with JSR 299 and play nicely together. I really hope that work gets done very quickly so that JSR 299 can go final for inclusion in EE6. It will be great to have SpringSource weigh in on 299, but hopefully they won't resort to childish name calling and dismissal.

Posted by Ryan de Laplante on June 09, 2009 at 10:08 PM PDT #

JSR 330 hass better annotation names than JSR 299.
Hopefully they go with the JSR 330 annotations

Posted by DK on June 10, 2009 at 02:14 AM PDT #

@DK - Of course, a key part of the complication is that the names (and eerything else) for JSR330 will not be final until JSR 330 goes final. SInce the two JSRs have different parties involved, the alignment is complicated. This would have been very easy to do if a proper DI for SE had been filed a while ago, when it became apparent it was needed...

Posted by Eduardo Pelegri-Llopart on June 10, 2009 at 02:19 AM PDT #

Something else to consider: Delay the release of Java EE6 until this is resolved. The few extra months delay would be much better than leaving Java EE6 without JSR 299 and we don't want to wait another 3 years to consider including it again.

I think they need to resolve more than just naming. I think there are fundamental differences in JSR 299 and 330's approach to DI, plus as Red Hat points out JSR 330 doesn't standardize enough and will lead to "vendor specific deployment descriptors" which is what Spring used as ammunition against Java EE from the beginning.

Posted by Ryan de Laplante on June 10, 2009 at 02:25 AM PDT #

@Ryan - Waiting a few extra months would probably be OK but, given the differences between 299 and 330 and the lack of JCP experience of the 330 leaders, it is not clear how many months that would take, nor what the final outcome would be. The Java EE 6 EG has been considering different alternatives; I hope they will be able to report on a plan "soon".

Posted by Eduardo Pelegri-Llopart on June 10, 2009 at 02:42 AM PDT #

Post a Comment:
Comments are closed for this entry.