W3C home > Mailing lists > Public > public-linked-json@w3.org > February 2013

Re: Problem with auto-generated fragment IDs for graph names

From: Pat Hayes <phayes@ihmc.us>
Date: Mon, 18 Feb 2013 23:07:16 -0600
Cc: "'Eric Prud'hommeaux'" <eric@w3.org>, "'Andy Seaborne'" <andy.seaborne@epimorphics.com>, <public-linked-json@w3.org>, "'RDF-WG'" <public-rdf-wg@w3.org>
Message-Id: <09AD4CE6-47B8-4462-A526-41C8E26F3150@ihmc.us>
To: "Markus Lanthaler" <markus.lanthaler@gmx.net>

On Feb 18, 2013, at 12:36 PM, Markus Lanthaler wrote:

> On Monday, February 18, 2013 7:08 PM, Eric Prud'hommeaux wrote:

...
> ...
>> I suspect that saying
>> 
>>    Within a dataset, a graph node label denotes a graph.
>>   Graph node labels may appear as subjects or objects in graphs.
>> 
>> would do the trick, but again, I don't understand what drove us from
>> "denotes" to "is paired with".
> 
> Neither do I.. I'm trying to find it out because I find it very confusing
> and inconsistent. Thanks for your patience. I know this all has already been
> debated to death but I wasn't able to find any arguments baking the decision
> up (apart from "the SPARQL WG is silent about it").

(Not to re-open the matter, but just to provide historical context.)

My recollection is that the issue was that there are already serious uses of quad stores in which the fourth quad field (which is the graph label when this is conceptualized as a dataset) is recognized as denoting something other than a graph. The two examples that were mentioned were the iuse of an IRI denoting an agent to "label" triples which are some content that was published or asserted by that agent, and the use of an IRI denoting a time interval (or something closely related to a time interval) to label triples which are understood to be true at the time. There may be other examples I do not recall. 

Some members of the WG were of the opinion that this did not prevent these IRIs from *also* denoting the graph, and dataset semantics based on this 'multiple-denotation' idea were proposed. I was (and still am) of the view that the basic idea of IRIs in RDF, and the 2004 normative semantics, are both incompatible with this idea of IRI's having multiple denotations, depending on "context". We did briefly consider extending RDF to explicitly allow contexts, but this is both a very large change to RDF and seems too complicated for practical use, so it didn't fly. So we are left with the strange situation that applications can use IRIs to "name" a graph which cannot denote that graph, because they are being used to denote something else. 

Pat

> 
> 
> 
> --
> Markus Lanthaler
> @markuslanthaler
> 
> 
> 

------------------------------------------------------------
IHMC                                     (850)434 8903 or (650)494 3973   
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
Received on Tuesday, 19 February 2013 05:07:51 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:39 GMT