RDF and mixed content

Now that we've clarified the major issues, and are just talking about 
the new drafts, I would not mind if we move to www-rdf-interest if you 
prefer.

Dave Beckett wrote:
> ...
> I'm not taking offence, just trying to work out if you have any new
> information.  This mixed content problem, I still can't work out what
> you, or others are seeing is the concern.

The mixed content issue was mostly a misunderstanding. It stems from the 
following sources:

  7. Eschew mixed content.

http://www.xml.com/pub/a/2002/10/30/rdf-friendly.html?page=last

It suggests instead of using inline mixed content, use a link to an 
element with mixed content.

Then, I see the RDDL spec, by two people with a long association with 
RDF, and it does the same thing:


<rddl:resource>
   <p id="rng-prose">A RelaxNG schema for the L language.</p>
   <rdf:description about="http://example.org/schemas/L.rng"
     rddl:title="Relax NG Schema">
     <rddl:nature rdf:resource="http://relaxng.org/ns/structure/1.0" />
     <rddl:purpose rdf:resource="http://www.rddl.org/purposes#validation" />
     <rddl:prose rdf:resource="#rng-prose" />
     </rdf:description>
</rddl:resource>

So then I read the RDF M&S specification to verify my impression that 
there is no way to do mixed content.

First I looked at the basic grammar and then the "abbreviated" grammar. 
Neither raised the possibility of literal XML elements (parsedType 
etc.). Now I see that later in the document there is a section called 
"Formal Grammar" which claims to summarize the grammar defined 
elsewhere, but actually extends it with extra features: "The complete 
BNF for RDF is reproduced here from previous sections." The title also 
suggests that this section formalizes the grammar defined informally 
elsewhere. But actually it seems to introduce a bunch of new syntax. But 
that's water under the bridge, the new drafts seem better.

 >...
> Maybe you'll be happy to know we have added typed literals to RDF and
> RDF/XML, so that the following is legal:
> 
>  <Class>
>      <Property rdf:datatype="&xsd;integer">5</Properyt>
>  </Class>
> 
> (the attribute value is a URI-reference, I'm abbreviating here).

Yes, that is progress. Also the handling of XML Literals is MUCH clearer 
in that specification.

That said, I see a new (to me) concept of forging a document with an 
rdf-wrapper. I do not think that this is not necessary. Just as the 
integer 5 does not have to be "wrapped" to be a value, the nodeset 
corresponding to an XML literal should not have to be wrapped. It is 
just a value. Or else you could think of them as graph nodes with 
identity. Either model is better than forging a string context for data 
that is necessarily already parsed by the time it is interpreted.

> See
>   Typed Literals
>   http://www.w3.org/TR/rdf-syntax-grammar/#section-Syntax-datatyped-literals
> 
> 
> Maybe the examples in the new syntax draft, out today, might help
> you?

Yes!

  Paul Prescod

Received on Wednesday, 13 November 2002 02:28:07 UTC