Re: why distinguish between simple and complex types? (issue 11): allowing internal and external ModelReferences on the same element

Dear Amit,

you're saying that "we need to allow redefinition/overwriting", yet you
present no use case that would actually require that, as opposed to
allowing both/all modelReferences to apply simultaneously. 

For example, when the type says something is an address, and the element
says it's a delivery address, these two annotations taken together do
not pose any harm that I can see. Am I missing something?

We need justification for features we'd like to introduce.

Best regards,

Jacek

On Tue, 2006-06-06 at 11:45 -0400, Amit Sheth @ LSDIS wrote:
> Jacek:
> 
> Here is a perspective on having both internal and external 
> ModelReferences on the same element.
> 
> If there are more than one ModelReferences for an
> element plus type (ComplexType/SimpleType), we need to be able to specifiy
> which one applies (hence the ability to identify precedence).
> [We need to allow redefinition/overwriting which is required when we 
> need to have
> semantic annotation of an element (ModelRef) wrt to more than one 
> ontologies,
> and redefinition/overwriting enables us to specify new annotation we wish to
> apply.]
> 
> Each ModelReference may have corresponding
> Schema Mapping. 
> (b1) I suggest we consider adequacy of "latest annotation applies".
> (b2) I do not believe we need to worry about mutual consistency between 
> two ModelReferences on an element
>      if they are wrt to different ontologies.
> (b3) The issue of consistency is important with respect to all 
> ModelReferences in a
> WSDL wrt to any one ontology (ontology models the world and has KR rigor 
> such as
> consistency of ontological specification), but this will need to be 
> handled by tools and developers.
> 
> Amit
> =====
> 
> Kunal, 
> 
> do you have any specific scenario where precedence rules would be
> useful? I feel that if the type says it is an Address, and the element
> that uses the type says it is DeliveryAddress, both do apply, right? 
> I don't really see how we could specify that DeliveryAddress applies
> more.
> 
> If there is a conflict, like the type says it is a "Mammal" and the
> element says it is a "Car", that would make an inconsistent (and
> invalid) SAWSDL document, and I don't think we should hide this problem
> by specifying that only Car applies for this particular use of what
> elsewhere would be Mammal.
> 
> So in a nutshell, I don't think we need precedence or resolution rules
> if we call inconsistent documents invalid.
> 
> Best regards,
> 
> Jacek
> 
> On Mon, 2006-06-05 at 17:25 -0400, Kunal Verma wrote:
> > Finally, allowing annotations for both elements and complexTypes begs
> > the question of which takes precedence when used together. As pointed
> > out by Laurent in
> > http://lists.w3.org/Archives/Public/public-ws-semann/2006May/0043, the
> > approach of giving the element annotation precedence over the type
> > annotation seems like the way to go.
> > 
> > "If some internal annotation exists for a complex type as well, any
> > "where used" annotation takes precedence over the internal one." 
> 
> 
>

Received on Monday, 12 June 2006 14:08:08 UTC