Re: OWL XML serialization

On 5 Mar 2008, at 13:53, Rinke Hoekstra wrote:

>
> Dear all,
>
> I just had a quick look at the OWL XML description and have some  
> comments on it.
>
> 1) What is the rationale behind using owl11xml:URI instead of  
> xml:id attributes on XML elements?

I would think that xml:id is inappropriate for several reasons:
	1) we aren't trying to identify elements; these are application  
level not XML level identifiers;
	2) the value of an xml:id is an NCName
		http://www.w3.org/TR/xml-id/#processing
	thus can't capture all URIs
	3) the ultimate type is "id" which may have some consequences for  
the PSVI.
	4) the uniqueness constraint is irrelevant so confusing

> It seems natural to adopt the standard way of identifying XML  
> elements.
[snip]

Well, my guess on what an XML weenie might say is that since we  
*aren't* identifying XML elements (but OWL classes, properties, etc.)  
that rather than being natural, it would be A Wrongness. Your inner  
XML weeniemeter might read differently :)

Cheers,
Bijan.

Received on Wednesday, 5 March 2008 18:21:19 UTC