Re: Possible solutions for ISSUE 97

O.k. Then, when you ask:

[[[
So if you are now saying that in order to support rdf:XMLLiterals, we
need to point out that the RDF graph produced by an RDFa parser would
have XML literals in, which *by definition* are in Exclusive
Canonicalised Form, then I can live with that.
]]]

this sounds fine, but I believe that, as an informative note, we should 
point out that this does _not_ necessarily mean that if the output of an 
RDFa processing is a particular serialization, then _that_ serialized 
format _must_ contain a canonical XML for XML Literals. Otherwise 
implementors may think otherwise.

So why is this fundamentally different from what I said in:

http://www.w3.org/mid/47E2287A.9000206@w3.org

[[[
Ie, *as far as the RDFa syntax document goes*, I think that the 
modifications to be done are purely editorial

- the SPARQL tests should be updated to add the xmlns namespaces 
somewhere on the top xml elements (ie, the <sup> in our Einstein examples)

- an informative note may have to be added to the syntax text warning 
implementers that they have to add the necessary namespaces

- maybe an extra test should be added to the suite that checks whether 
the xml:lang attribute has been properly added to an XMLLiteral output, 
too.
]]]

Ivan


Mark Birbeck wrote:
> Hi Ivan,
> 
> On 20/03/2008, Ivan Herman <ivan@w3.org> wrote:
>> Mark,
>>
>>  I still need to understand something in your arguments.
>>
>>  Let us take this RDFa:
>>
>>  <html xmlns="http://www.w3.org/1999/xhtml"
>>        xmlns:q="http://b.b.b">
>>  <head>
>>  </head>
>>  <body>
>>    <div about="http://a.b.c" property="q:r"><span>ABC</span></div>
>>  </body>
>>  </html>
>>
>>  Say that my implementation generates the following RDF/XML.
>>
>>  <rdf:RDF xmlns:rdf="....">
>>     <rdf:Description rdf:about="http://a.b.c">
>>       <q:r rdf:parseType="Literal"><span
>>  xmlns='http://www.w3.org/1999/xhtml'>ABC</span></q:r>
>>     </rdf:Description>
>>  </rdf:RDF>
>>
>>  Note that the content of the <q:r> is _not_ canonicalized XML, because I
>>  use xmlns='...' instead of xmlns="...". So my questions
>>
>>  1. Is this RDF/XML portion, in your view, valid RDF/XML?
> 
> Yes, it is, since RDF/XML as a serialisation of an RDF graph does not
> prohibit you from using apostrophes.
> 
> 
>>  2. Is this output conformant with the RDFa syntax definition?
> 
> In what way? I'm having trouble seeing what RDFa has to say about
> RDF/XML serialisations at all. Or N3, or Turtle, or any other
> serialisation of an RDF graph.
> 
> We went to a lot of trouble to ensure that RDFa was all about RDF, and
> therefore independent of any particular serialisation. (Hence defining
> everything in terms of an RDF graph.)
> 
> 
>>  My claim is that the answer on both questions are 'yes'. And, in my
>>  view, the RDFa syntax document should make that clear.
> 
> Although the answer to the first question ("is this valid RDF/XML") is
> 'yes', that's not the same question as "is this a correct
> serialisation of the RDFa". To go from RDFa to RDF/XML you have to go:
> 
>   RDFa --> RDF --> RDF/XML
> 
> which means you must have created an abstract graph (the RDF in the
> middle), which in turn means you must have canonicalised the XML. So
> to produce the RDF/XML that you have, with a literal that is
> non-canonicalised, you must have substituted the quotes for
> apostrophes! Your implementation must have de-canonicalised. :)
> 
> (But that doesn't matter, because if you parser that and convert it to
> an abstract graph, it will get normalised again; it's just an odd
> thing to do.)
> 
> To say again, just because you can represent something as valid
> RDF/XML, doesn't mean therefore that the abstract graph can take
> exactly the same form.
> 
> Regards,
> 
> Mark
> 

-- 

Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf

Received on Thursday, 20 March 2008 15:09:05 UTC