On Aug 15, 2007, at 2:20 AM, Aristedes Maniatis wrote:
If we do expect people to use it in production, why not release it
as 2.5 or 3.0? Sure JPA isn't complete, but lifecycle events and
lots of other things are. I'm using it in production for that
reason, but we can't expect everyone to do that for what sounds
like an alpha release.
I see your point here, and I agree that it would be ideal if we could
do it realistically. What's holding us from doing this is that the
new feature development is not serial, but often parallel. I.e. we do
"Y" before we finish "X", resulting in a bunch of loose ends that
can't be considered production quality. In case of 3.0 some loose
ends are:
* embeddable support (half way there)
* modeler support for lifecycle callbacks
* documentation and tutorials for the partial EJBQL ("partial" is
also a confusing part as we can't clearly tell the users what parts
of the syntax should be avoided)
* some other things I can't remember now.
Switching to a serial development mode would place extra burden on
developers (likely resulting in early branching). I don't know -
maybe we should return to this discussion once we start TCK testing
(I haven't yet) and see where we are in terms of completeness.
There is no rush to release, but is the goal: full JPA compliance?
Yes, unless we hit the wall with TCK testing.
How will we know when we have arrived? (especially with the JCK
issues).
We have access to JSR-220 TCK (I do at the moment; anybody else can
also sign an NDA and ask Geir to give it to them).
Andrus
