Antonio Goncalves wrote on 07/16/11 12:31 AM:
> I know EE 7 is time driven and we can't put all the features into it.
> Configuration is an important topic, but as you say Bill, not as easy as it
> seems because it has to fit with the existing EE deployment descriptor. Why not
> follow the Bean Validation 1.0 spec ? In Apendix C of the spec, the EG decided
> to make a proposal for method-level validation. It even starts with the
> following sentences :
>
> "This proposition has not been integrated into the core specification and is not
> part of it. It remains here for archaeological
> purposes and will be seriously considered for a future revision of this
> specification."
>
> Why not follow this example ? The EE 7 EG could start talking about it, write a
> proposal, either we have time and it goes into the spec, or we don't and it
> stays in the Apendix. The nice thing with that, it that we put into writing what
> we discussed. This way, if configuration has to wait EE 8, at least we will have
> something to start with.
I'm hoping our new approach to transparency will allow us to do a better job
of this. I don't think we need such a proposal as part of the spec; we can
put it in the java.net project so that others can review it and so that it's
available as a starting point for the next expert group.
> This being said, I think configuration is a more important topic that PAAS/SAAS
> (which I find a bit too young, but I've already said that) and should make its
> way into EE 7.
Just saying "configuration" is too imprecise. We already have configuration.
Changing it so that you use a different syntax to do the same things doesn't
seem important. You need to describe what problem needs to be solved that's
more important to the success of the Java EE platform than providing PaaS
support.