W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > January to March 2009


From: Bijan Parsia <bparsia@cs.manchester.ac.uk>
Date: Tue, 31 Mar 2009 22:33:11 +0100
Message-Id: <E2836CD7-3541-4EB7-AD0A-257456845B70@cs.manchester.ac.uk>
To: Lee Feigenbaum <lee@thefigtrees.net>, RDF Data Access Working Group <public-rdf-dawg@w3.org>
On 31 Mar 2009, at 22:18, Lee Feigenbaum wrote:
> I have two concerns with the SPARQLX proposed feature:
> 1) Is it currently implemented anywhere? Looking through the old  
> discussions it sounds like Kendall was going to be implementing it;  
> do you know if he did? do you know if anyone else did?

What counts as "implementation" here? I plan to provide an XSLT and an  
XSugar dual grammar implementations for a converged XML Schema. Other  
tools will then work as is since they can consume the normal SPARQL  

> 2) As the query language evolves, so too will the XML syntax have to  
> evolve. This puts a small burden on any people extending the  
> language in the future, but more than that, what's the latest state- 
> of-the-art with respect to evolving XML schemas? Does extending a  
> schema as the language extends break backwards compatibility for  
> tools that validate against an older schema? (As you can probably  
> tell, I'm not much of an XML-head, so apologies if this question is  
> non-sensical.)

There's two basic strategies:
	1) Same as the current grammar. When you extend, the old set of  
queries remain (in general) valid, but queries using new features are  
not legal under the old grammar.
	2) Use extensibility points. XML Schema has a lot of facilities that  
let you control how you can add or remove things. Far too many to go  
into here. But the obvious one to note is "derivation by extension"  
which lets you take an old element type and extend its content model  
(e.g., to add extra stuff at the end).

> (FWIW, while these are concerns, they're both mild concerns. On the  
> flip side, I've been asked about XML serializations of SPARQL  
> repeatedly in the past, so I think there is sufficient demand for it.)
> To me, the RDF syntax proposed by Holger has the benefit that there  
> is one implementation of it and the start of a 'specification', but  
> I haven't heard much clamor elsewhere for an RDF serialization.

Also, it just doesn't meet the toolchain requirement. See my video for  

(more vids coming :))

Received on Tuesday, 31 March 2009 21:33:48 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:00:56 UTC