In a very geeky blog, BEA says the same about JDO that I've been thinking would be the case for a while:

BEA explained their no vote saying they 'don't see how having another release of JDO, whose market acceptance is essentially constrained to use with object databases, can be explained in conjunction to similar enhancements being made in the EJB3 expert group.'

I just can't see a good reason to use it. I think in rdbms-ese. No real good reason for me to serialize whole objects to a database; that seems to blast the whole advantage I have by using the rdbms in the first place. Anyhow, looks like BEA is saying the same thing, in this case wrt a new version of JDO and why they voted against releasing the spec.