Section 6.1.2 - Do we need clarification?

[tviegut]

For paragraph in Section 6.1.2:

The new attributes or associations can be added to the new extension class and they will be inherited by the standard CIM class and in effect appearing as new properties of the standard class.

Tonja Kostic commented:

This may be visible in the profiles, but for the doc generation and validation, I’m not sure it will be easy to implement and fully support, so you’ll be on your own.

More importantly:
This will invalidate any XML Schema profile generation, because XSD does not support MI, and profiling tools have to find a way how to deal with this. CIMTool might have a trick (to merge into one class attributes from multiple classes with the same name), but this is nowhere documented.

Kendal commented:

I think we agreed that profiles would not use MI. Profiles allow inheritance flattening and thus force the flattening of extensions into the the original class. In UML profiles this then requires a namespace applied to the flattened properties, which was the motivation for the "extra" stuff beyond package namespaces in the namespace discussion. This allows traceability through the various profile layer models. I don't think traceability directly to the original canonical model is sufficient, we need traceability through the various stages of profiling back to the canonical model, especially if any profiles layers add semantics as some WG16 profiles have done.