Chiusano Joseph wrote:
>Farrukh,
>
>I think CCOWL is a great idea. I'm also thinking that we can have the
>best of both worlds here - that is, we don't need to halt our current
>CCRIM work in order to pursue incorporation of semantic technologies.
>The reason I say this is that my understanding is that OWL would apply
>to the assembly functionality in the CCTS spec, which is out of scope of
>the CCRIM effort anyway. Basic registration and maintenance of Core
>Components and their associated entities in the registry would be
>covered by the base registry functionality for handling RegistryObjects.
>
>Does that sound good?
>
>
Joe,
Incremental progress is always a good idea IMO. Indeed we could define a
Technical Note binding CCTS to V3 RIM and then map that work to a new
Technical Note binding CCTS to OWL and expecting that there will be a
normative mapping of OWL within ebXML Registry in version 4.
I do want to emphasize though that OWL is not just applicable for
assembly but also for expression of CCTS in XML. In fact as I look at
some of the issues you identified in expressing CCTS in RIM (for
example the Slot limitations), I notice that in an OWL expression, those
limitations simply go away. For example, RIM when expressed in OWL do
not even need to have the notion of a Slot class since OWL has an
inherent ability to express slots or dynamic attributes.
I am beginning to think that we should not make major changes to RIM to
fix the various limitations identified in V3 (e.g. Slot fixes) and
instead focus on fixing them in V4 when we move toward OWL as an
expression syntax for RIM.
As for assembly, I have raised the issue privately with David Webber and
will do so on the CAM mailing list as well, that OWL seems to be a
better fit for CAM assembly expression than a custom CAM schema based on
XML Schema. The reasons are the very same reasons why CAM finds XML
Schema to be inadequate for expressing the rich semantics of assembly
rules and constraints. OWL gives CAM a richer expression of these
assembly rules. But that is for a different thread on CAM.
In summary, the last thing I want is for us to destabilize the CCRIM
work. Instead I want to show the connection between this work and the
Semantic Content Management work and suggest we do the CCRIM work for V3
based on knowledge of the future directions of V4 toward direct OWL
support within RIM.
--
Regards,
Farrukh
