Re: [RDFa] rdf:XMLLiteral (was RE: Missing issue on the list: identification of RDFa content)

Hi Dan,

This is a good question, but it's much the same as the SPARQL issue
raised by Ivan; since RDF Concepts is quite narrow about the notion of
equality then we already have the kinds of problems you are alluding
to. For example, in Ivan's FOAF data (in RDF/XML and in RDFa) he
sometimes has his name with no language and sometimes with a language;
this means that as far as RDF Concepts is concerned two exactly equal
strings will not be equal.

SPARQL gets round this by adding some extra functions to make
comparisons, but my point is that we already have problems with data
formats, that are not solved by RDFa using plain literals rather than
XML literals.

I mention this because, as you rightly point out, any vocabulary that
uses rdfs:Literal is *already* allowing both rdf:XMLLiterals and plain
literals (and any other data type too). So any other RDF serialisation
could give rise to XML in a FOAF name, not just RDFa.

So for me, the question remains, what do we *lose* on the RDF side,
given that on the XHTML side we *gain* the ability to mark-up text
without having to resort to using the datatype attribute and the RDF
namespace.

Regards,

Mark

On 18/03/07, Dan Brickley <danbri@danbri.org> wrote:
> Mark Birbeck wrote:
> >
> > Hi Elias/Ian,
> >
> > I'm afraid I'm missing from this discussion, first what we *lose* by
> > using rdf:XMLLiteral, and second, some clear-cut explanation of why
> > plain literals are *logically* the correct default, rather than just
> > simply someone's 'preference'.
>
> One brief but hopefully simple point: if RDFa generates literals that
> are typed XMLLiteral, ... RDFa document authors need to choose RDF
> vocabularies whose properties have that has a range.
>
> Actually I'm not sure. They certainly need to consider the range.
>
> http://www.w3.org/TR/rdf-schema/#ch_xmlliteral
>
> "The class rdf:XMLLiteral is the class of XML literal values.
> rdf:XMLLiteral is an instance of rdfs:Datatype and a subclass of
> rdfs:Literal."
>
> If we define a property to have a range rdfs:Literal, and it is
> generally used with plain literals... does anything get tricky if we
> start using it with rdf:XMLLiteral?
>
> I have to admit, to date, I had assumed without scrutiny that this was
> problematic. I guess I had been confusing the superclass rdfs:Literal
> with the notion of a "plain Literal",
> http://www.w3.org/TR/2004/REC-rdf-concepts-20040210/#dfn-plain-literal
> ... but it seems (unless I'm missing something; sorry I forget the
> design discussions! it was a while back now...) ...seems that we don't
> define a class for plain literals.
>
> So, for example, we say foaf:name has a range
> http://www.w3.org/2000/01/rdf-schema#Literal
>
> I had previously thought this made use of XMLLiteral in names
> problematic, eg. for ruby markup in names. But perhaps not.
>
> A question (for the OWL folk here): if we have a property sometimes
> taking plain literals as values, and sometimes taking XMLLiteral, ...
> does this put the property (and hence vocab) into OWL Full?
>
> cheers,
>
> Dan
>
>
>
>
>
>
>
>
>


-- 
  Mark Birbeck, formsPlayer

  mark.birbeck@x-port.net | +44 (0) 20 7689 9232
  http://www.formsPlayer.com | http://internet-apps.blogspot.com

  standards. innovation.

Received on Monday, 19 March 2007 01:58:19 UTC