Re: 2001-09-07#6: ns qualified parseType values

dehora wrote:
> 
[...]
> The worst case is
> an old parser treating 'rdf:Resource' as 'Literal': I acknowledge that
> interop would fail in this case where receiving software has not been
> upgraded, but that would a software defect rather than a spec defect.
> Asking implementers to inspect parseType attribute value strings
> containing ':' and check against the qualified name for the RDF
> namespace is a simple enough task, certainly no harder that checking
the
> attribute itself for namespace qualification, which we're already
asking
> people to do.

Dan Connolly :

It's a small change, but not one that I think can be reasonably
read into the RDF 1.0 spec. We're not designing RDF 1.0; we're
clarifying the spec, right?
>>

I don't draw a useful distinction between a clarification and a redesign
if that's whats needed. I figure if we can go as far as producing a
model theory to clarify the spec, we can certainly add some words about
using parseType to qualify literals.


Dan Connolly :

If/when we get around to designing, I'll suggest that
we replace rdf:parseType with xsi:type (from XML Schema) wholesale.
>>

Sure, you can do that. Or maybe find a way to extend rdf:parseType from
xsi:type and stay somewhat backward compatible (could that be done
inside RDFS' mandate now? Danbri??). In the meantime people will still
have to tunnel type information about XML-RDF Literals in some form or
another. So we might as well offer them an upgrade path.

Bill

Received on Wednesday, 19 September 2001 09:32:55 UTC