On Sep 9, 2009, at 10:54 AM, Kevan Miller wrote:
Personally, I think I've been guilty of allowing spec instabilities to affect my judgement on community readiness (or at least affecting how much I pushed the community to consider graduation). IMO, I think that was wrong. I should not be allowing external factors to be affecting my community evaluation.
It is rather understandable. For a while there the spec was wavering between being something that could be implemented standalone and something that was completely tied into the app server. It certainly does make it hard to know if TLP or subproject is the right choice which makes pushing for graduation a tricky subject. The spec settled as a little both, so either TLP or subproject could work using the technology described by the spec as reference.
That said, the specs have always said EJB couldn't be a standalone technology, yet OpenEJB did exactly that for it's entire existence. After existing 10 years and 4 spec versions we finally got the concept of an EJB container usable in a Java SE environment, outside an app server, officially added to the specs. So there is something to be said for blazing your own trail.
-David
