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

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

From: Dave Reynolds <dave.e.reynolds@gmail.com>
Date: Tue, 19 Feb 2013 09:37:08 +0000
Message-ID: <512347C4.9020205@gmail.com>
To: Pat Hayes <phayes@ihmc.us>
CC: Eric Prud'hommeaux <eric@w3.org>, Markus Lanthaler <markus.lanthaler@gmx.net>, 'Andy Seaborne' <andy.seaborne@epimorphics.com>, public-linked-json@w3.org, 'RDF-WG' <public-rdf-wg@w3.org>
[Sorry if commenting here is inappropriate and no doubt this will bounce 
off some of the lists. Apologies if I'm way off the mark here, not been 
following the discussion in detail ...]

I wonder if you have the option to approach this from the vocabulary end 
rather than than the general semantics of RDF.

Surely in RDF we rarely know what an IRI denotes. We determine that (or 
at least constrain it) from the statements we make about that IRI.

On it's own we know nothing of what is denoted by:

       <http://www.amberdown.net/rdf/foaf.rdf#der>

But in a graph where we see:

     <http://www.amberdown.net/rdf/foaf.rdf#der> a foaf:Person;
          foaf:name "Dave Reynolds" .

then we deduce that it denotes a person.

So if the presenting use case is about signing graphs then the hard and 
important part is how you define this x:signature predicate.

You might have something like:

     x:signature a owl:DatatypeProperty;
         rdfs:domain x:GraphInDataset;
         rdfs:range  xsd:hexBinary;
         rdfs:comment """
            Indicates that there is a Dataset in which the subject is 
the label of a graph which when cannonicalized using algorithm A, 
serialized using algorithm B and then signed using algorithm C generates 
the signature given as the value of this predicate.
If this predicate occurs within the default graph of a dataset, and if 
that dataset contains a graph labelled with the subject IRI then a 
processor SHOULD assume that the graph so labelled is indeed supposed to 
have the given signature. In other situations the processor may only 
know that there is a graph in existence with this signature but may not 
have a means to obtain it.
         """ .

Not saying that's a *good* definition, and there are certainly lots of 
alternatives that would have different consequences. My point is that in 
order to be able to state signatures of graphs in JSON-LD you need some 
such definition and it's not clear to me that you need a change in RDF 
semantics to make that possible.

[I'll shut up again now.]

Dave

"IRIs mean whatever I want them to mean ..."
Received on Tuesday, 19 February 2013 09:37:40 GMT

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