|Re: [stp-newsgroup] SCA v0.95 EMF model contribution? [message #587235]
||Wed, 23 August 2006 02:04
| Oisin Hurley
Registered: July 2009
> We are updating the schemas and models now to react to the 0.95 |
> (looking at 0.96 now as well). For the most part the changes are
> aligned with the direction that we wanted.
> That is that Composite has been added as a first class type. The
> big change is that Module, EntryPoint, ExternalReference, and
> ModuleComponent are removed. The main concepts still in the spec
> are the following:
> Composite is an implementation and can be set as the implementation
> of a component. EntryPoints have been replaced with Service and
> ExternalService has been replaced with Reference. Composites
> support "includes" semantics to enable teams to divide the content
> of a Composite into more manageable bits. When a Composite is
> included in another Composite all of its contents are "absorbed"
> into the definition of the containing Composite. Therefore, the
> Composite is nothing more than a container in this situation.
> The spec hasn't changed Subsystem too much except to remove
> ModuleComponent and SystemWire. However, there is still a great
> deal of churn around the deployment model in SCA so we are going to
> avoid this area of the model for now until the specification
> settles on an approach.
Agreed on the deployment churn - this is something that will need to
in the spec. As you say the changes are quite compatible with the
current Core approach,
but that doesn't necessarily mean it will be trivial to move the
code :) Do you think
there will be a big impact on the core code as it exists right now?
Is the transition
to a 0.95++ version of the spec a bumpy one from the implementation
Is it feasible to construct a roadmap from the current Core state to
a 0.95++ compliant
state? - the reason I ask is that if you can put a roadmap together
then we may be
able to distribute the task of making the transition happen.
Powered by FUDForum
. Page generated in 0.31149 seconds