W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > January to March 2011

Re: Service or graph store naming.

From: Andy Seaborne <andy.seaborne@epimorphics.com>
Date: Mon, 14 Feb 2011 22:10:36 +0000
Message-ID: <4D59A85C.7030401@epimorphics.com>
To: Gregory Williams <greg@evilfunhouse.com>
CC: Axel Polleres <axel.polleres@deri.org>, SPARQL Working Group <public-rdf-dawg@w3.org>


On 14/02/11 21:10, Gregory Williams wrote:
> On Feb 14, 2011, at 4:04 PM, Andy Seaborne wrote:
>
>> The service is replicated (to spread load, better latency by being
>> close, etc) - relying on rotating DNS would mean the path at all
>> points would need to be the same.  Think of it more like download
>> mirroring.
>>
>> This means there is <myService>  a sd:Service ; sd:url</sparql2>
>> ,</sparql2>
>
> I assume those were meant to be two different urls?

Yes.

>
>> :sd:url has domain service and range endpoint.
>>
>> maybe: """ 3.2.1 sd:url
>>
>> The service URL of an sd:Service supporting the SparqlQuery
>> interface. The object of the sd:url property is a URI reference
>> """ ==> """ The service URL of an sd:Service supporting the
>> SparqlQuery interface. The object of the sd:url property is a URL
>> for a service endpoint. """
>
> And the conformance would then include something along the lines of:
>
>  The RDF content returned from dereferencing a service URL<U>  must
> contain one and only one triple matching: ?service sd:url<U>  .
>
> ?

How about making sd:url an inverse functional property?

>
> .greg
>

	Andy
Received on Monday, 14 February 2011 22:11:15 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:45 GMT