Re: The URI of a RDDL "nature"

noah_mendelsohn@us.ibm.com wrote:
> ...
> Well, I think that the nature of an XML document may well be  
> determined,
> or at least bounded, by the >QName< of the root element.  I don't  
> think
> the namespace does it. Consider a single namespace that includes  
> two or
> more element QNames both of which were designed to be used as root
> elements.  For example, I might in the same namespace (shown with  
> prefix
> ns:) have:
>
>         <ns:purchaseOrder>
>         ....
>         </ns:purchaseOrder>
>
> and also
>
>         <ns:invoice>
>         ....
>         </ns:invoice>
>
> Surely it's wrong to say that the nature of these documents is  
> determined
> by their namespaces.  One's a purchase order, the other an  
> invoice.  Both
> are in the same namespace.  I think RDDL should be capable of  
> capturing
> these separate natures.

RDDL can. If we consider the rddl:nature as determining an rdf:type  
then this is the same as type inheritance.

e.g. one can say that the "nature" of all documents which match the  
pattern

<ns:*>
</ns:*>

is that they all have root elements which are qualified by the  
namespace <ns:>

Documents with a root: <ns:purchaseOrder> are a sub-class of these  
just as are documents with a root <ns:invoice>

Moreover the author of the namespace document is asserting the  
rddl:nature so this can be as specific as is needed. Using the root  
element namespace is meant to be a guide.

If we can't use the root element namespace as a guide we just need to  
tell people (i.e. namespace document authors) a better way to  
determine what the nature of a resource ought be.

Jonathan

Received on Wednesday, 18 January 2006 00:07:32 UTC