W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > April to June 2010

Re: Service Description document

From: Andy Seaborne <andy.seaborne@talis.com>
Date: Tue, 27 Apr 2010 09:15:52 +0100
Message-ID: <4BD69D38.3000404@talis.com>
To: Gregory Williams <greg@evilfunhouse.com>
CC: SPARQL Working Group WG <public-rdf-dawg@w3.org>


On 27/04/2010 3:24 AM, Gregory Williams wrote:
> The second issue is one involving the naming of named graphs in a dataset description. Sandro had brought this up at the F2F, and I hope he can weigh in on it. Basically, I had the modeling in the document like this:
>
> [] a sd:Dataset ;
>      sd:namedGraph [
>          sd:name<graph-uri>  ;
>          sd:graph [ ... graph description goes here ... ] ;
>      ]
>
> I believe Sandro's concern was the use of the URI<graph-uri>  since we're not actually talking about the resource, but instead a name that we've locally given to the graph (hopefully he'll correct me if I'm mis-characterizing this). Instead, he suggested that it should be "graph-uri"^^xsd:anyURI. If we adopt this change then the document should clarify the rdfs:range of the sd:name property and the Example section should be updated accordingly.

I'm not sure this is helpful.  The named graph paper uses URIs, as do 
TriG and NQuads, as well as SPARQL query. They don't use xsd:anyURI.

The "local name" idea does not seem to me to get round the issue for 
xsd:anyURI as it's still associating a global identifier with the graph, 
even if it's not web-naming it.  The use in a web document means we 
still have a global reference to the thing.

	Andy
Received on Tuesday, 27 April 2010 08:16:23 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:42 GMT