RE: Update to XLink -> RDF mapping

Hi all,

In yesterday's XLink working group teleconference, the
topic of the mapping of XLinks to RDF came up. The
question, in essence, is whether there is enough interest
in this work to justify completing it.

So far, there has been very little traffic about the
mapping on the RDF interest list. Nor has there been any significant
off-list email. (I recall receiving one
message saying that the sender thought the work was
probably important, but they had no time to look at it.
That was in response to the first version of the mapping
posted 5/13.)

So, if you want this work to continue, please express an
interest.  The most recent version is attached. (Eve,
I still owe you some cleanups on the MUST/SHOULD/MAY
points you flagged as possibly ambiguous. Also, I might
work in an XSLT example based on the stuff I sent to the
RDF-interest list on the 11'th.)

 <<xlink2rdf2.html>>  <<xlink2rdf-fig3.gif>>  <<xlink2rdf-fig1.gif>>  
<<xlink2rdf-fig2.gif>> 

Regards,
Ron Daniel Jr.
Metacode Technologies, Inc.
139 Townsend Street, Suite 100
San Francisco, CA  94107
415.836.7813 fax 415.222.0150 
rdaniel@metacode.com





> -----Original Message-----
> From:	Ron Daniel 
> Sent:	Tuesday, July 11, 2000 11:40 AM
> To:	Dan Connolly
> Cc:	www-rdf-interest@w3.org
> Subject:	RE: Update to XLink -> RDF mapping
> 
> Hi Dan,
> 
> Thought I had sent a response earlier but don't see
> it, so here's the old response plus some new stuff...
> 
> 
> Dan Connolly said:
> 
> 	an editorial nit: it says both
> 
> 	| The
> 	| starting Resource of the simple link SHALL be mapped to the
> subject of
> 	the RDF statement.
> 
> 	and
> 
> 	| The ending Resource of the simple link SHALL be mapped to the
> subject
> 	of the RDF statement.
> 
> 	I presume you meant 'object' rather than 'subject' in the latter.
> 
> Yep, thanks for catching that. I've updated it in the
> XML source of the NOTE and it will go out with the
> next version.
> 
> 	Anyway... I got about 1/10th of the way thru implementing it
> 	as an XSLT transformation. I was able to convert your first
> 	example:
> 
> 	... In a <x:extRef xlink:type="simple"
> 	      xlink:href="http://www.foo.com/papers/crops.txt"
> 	      xlink:arcrole="http://links.org/namespace/cite"
> 	      xlink:role="http://links.org/namespace/screed"
> 	   >recent paper</x:extRef>, Dr. Taylor assumes that ...
> 
> 	to
> 
> 	<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
> 	xmlns:xlink="http://www.w3.org/1999/xlink">
> 	  <oops-predicate rdf:about="#xpointer-for-N8"
> 	  rdf:resource="http://www.foo.com/papers/crops.txt"
> 	  xmlns="http://links.org/namespace/cite"
> 	  xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" />
> 
> 	  <rdf:type rdf:about="http://www.foo.com/papers/crops.txt"
> 	  rdf:resource="http://links.org/namespace/screed" />
> 
> 	  <rdf:Class rdf:about="http://links.org/namespace/screed" />
> 	</rdf:RDF>
> 
> 	As you can see, I ran into some trouble already:
> 
> 		* how do I compute a namespace name and a local name from
> 		the predicate http://links.org/namespace/cite? I suppose
> 		l:cite xmlns:l='http://links.org/namespace/' would work,
> 		but I couln't figure out how to implement that with XSLT ;-)
> 
> 		But I don't see a general purpose algorithm, especially
> 		since URIs need not end with XML name characters
> 		(e.g. http://example.org/foo/).
> 
> 		This is a generic RDF problem, not a problem with this
> 		xlink->rdf mapping; i.e. it's an issue with serializing
> 		RDF models in the general case.
> 
> To be perfectly honest, I had been assuming that the
> harvesting would happen through something like XSLT extension
> functions that directly poked statements into an RDF store,
> rather than converting to RDF's serialization syntax.
> How to serialize a RDF model that never kept info about
> the namespace URI vs. local part is, as you point out,
> a thorny issue for all RDF implementations.
> 
> But your sample motivated me to actually try and write
> a proof-of-concept. Here's a XSLT stylesheet and a
> mockup of the extension function. Barring errors, I think
> the stylesheet handles all the cases with simple links.
> I have not done anything with extended links yet.
> 
> This particular stylesheet works with saxon's version of
> extension functions. It should not be hard to make it
> work with XT but I have not tried.
> 
> (BTW - people who see 'Radix.java' should not get
> their hopes up. This is nothing more than one call.)
> 
>  << File: map_simple.xsl >>  << File: Radix.java >>  << File: ex1.xml >> 
> 
> 
> Regards,
> Ron Daniel Jr.
> Metacode Technologies, Inc.
> 139 Townsend Street, Suite 100
> San Francisco, CA  94107
> 415.836.7813 fax 415.222.0150 
> rdaniel@metacode.com
> 
> 
>