In short, future-proofing attempts would be a waste of time.
On the other hand, making the new HornetQ journal based objectstore
implementation a user configurable option would have a huge performance
benefit for some scenarios. Likewise making the JTA/JTS choice
selectable - just having an ORB present does not necessarily mean a user
needs distributed tx and the associated performance penalty for all tx.
Put the limited time available into working on things that will have a
benefit, not things that may.
Ok I think you made it clear that it hasn't been designed yet.
We are supposed to try and keep the domain model compatible for as long as we can so
it's never a bad idea to ask what the future plans likely are, if they are known.
Imagine if they did exist and we never asked. We'd look pretty foolish if things
didn't work out!