On Tue, 9 Nov 2004 19:08:13 +0100, Laurent Le Meur
<[EMAIL PROTECTED]> wrote:
> 
> ATFU, http://www.xml.com/pub/a/2002/10/30/rdf-friendly.html gives more 
> information
> about what is being RDF-friendly.

Thanks, good pointer.

> But the modifications could maybe be smaller -> for example making the 
> globally
> unique "id" an attribute of <entry> (it can be mapped to an rdf:about 
> attribute with
> a simple xslt), defining the Atom-core entry elements as simple RDF-like 
> elements (no
> mixed content), and allowing any #other# namespaced elements in <entry>.

Hmm,  XSLT can take the <id> element content and put it into an
rdf:about attribute, but I think the aim here would be to avoid
needing XSLT if at all possible. It's not a typical use of a URI, but
if <id> was left as it is then it could be defined as an inverse
functional property when used with RDF.

> But: RDF is about metadata after all. The content part of the entry has 
> nothing to do
> with RDF. How would you make the separation?

Data is data. The content is a representation of the resource, there's
no reason that shouldn't be part of a Resource Description Framework
model. I've not followed the content discussion fully recently, but
I'm sure there will be a suitable datatype available in RDF.

In passing I think it's worth noting that RDF/XML literals changed
quite a bit after the original RSS 1.0 spec came out, and the usual
approach to content there (as <content:encoded>) shares some of the
problems associated with character escaping in RSS 2.0.  If there ever
is a revision of RSS 1.0 then it might well deprecate
<content:encoded> in favour of using RDF/XML's XMLLiterals with true
XML content (e.g. XHTML). This could lead to an element
<content:unencoded>...

Cheers,
Danny.

[1] http://www.w3.org/TR/rdf-syntax-grammar/#section-Syntax-XML-literals

-- 

http://dannyayers.com

Reply via email to