Re: Atom 1.0 + RDFa 1.1

On Nov 15, 2010, at 13:35 , Toby Inkster wrote:

> On Mon, 2010-11-15 at 13:02 +0100, Ivan Herman wrote:
>> Actually, this may be an idea on the specification level, too. Why
>> don't we specify these things as a virtual DOM->DOM transformation in
>> the document, too? Ie, let us not touch the formal processing steps of
>> RDFa Core; rather, any host language could be defined by such
>> transformations (and possibly a default profile). That would suggest a
>> much cleaner separation specification-wise.
> 
> I'd be happy with this provided that it doesn't lead to implementations
> being required to support XSLT. XSLT (2.0 in particular) is a pretty
> heavy-weight dependency.

Sorry if I was not clear. Of course we do not want a reference to XSLT, let alone a dependence. God forbid:-)


> 
> Perhaps something along the lines of: "Host language specifications
> SHOULD provide an XSLT transformation that may be used to transform a
> host language document into a document capable of being parsed as RDFa
> Core 1.1. Processing software MAY use this as part of a pre-processing
> step to handle the host language."
> 

Even less. I would scrap any reference to XSLT and just say that they should specify how a DOM (or XML Infoset?) may have to be transformed before passing it to the RDFa Core processing.


> The only problem with this -- and pre-processing in general -- is the
> XMLLiteral edge case. e.g.
> 
>  <feed xmlns="http://www.w3.org/2005/Atom"
>    property="http://example.com/xml"
>    datatype="http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral"
>> <entry /></feed>
> 
> Should result in the following triple:
> 
>  <> <http://example.com/xml> "<entry xmlns=\"http://www.w3.org/2005/Atom\"></entry>"^^rdf:XMLLiteral .
> 
> But if a pre-processor were used, might result in:
> 
>  <> <http://example.com/xml> "<entry xmlns=\"http://www.w3.org/2005/Atom\" typeof=\"\"></entry>"^^rdf:XMLLiteral .
> 
> The XSLT needs to be pretty clever to avoid things like this.
> 

And that is an XSLT problem that we should not deal with.

My implementation has some framework where anybody (even the caller through some sort of call backs) can add his/her transformation; this is simply a python function that gets a reference to the DOM representation of the input file and can be massaged by the function using DOM calls in Python. Ie, this implementation is oblivious to the serialization problem of the XML Infoset.

I guess an ATOM+RDFa description would simply describe that a transformation should be applied whereby each <feed> element gets a @typeof="" attribute before processing it through RDFa Core. 

Ivan


> -- 
> Toby A Inkster
> <mailto:mail@tobyinkster.co.uk>
> <http://tobyinkster.co.uk>
> 


----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf