Re: How do RDF and Formal Logic fit together?

[Seth Russell]

>
> There are two ways that I know which would allow us to extend RDF into a
> quadruple so that it has some context mechanism.  One way is to use a
> statement Id, and the other is to use a context uri.   The more I think
> about this,  the more I think that the context uri would be best.
> Seralizing the context uri could be done like this:
>
> <rdf  context="http://robustai.net/mentography/">
>     < ..... a batch of xml\rdf seralized statemets ... >
> </rdf>
>
> RDF 2.0 compliant parsers would then include the context uri as the fourth
> element of each quadrapule; if it were not present  they might substitute
> the document url.   How else are we going to searlize in XML\RDF 2.0 the
> contexts that N3 and Semenglish can express ?
>
> I know this is a departure from what I suggested earlier, but it has the
> advantage of not needing to come up with a scheme for making a globally
> unique statement IDs, it's far easire to write, and it gives us a place to
> put a context uri, which we needed anyway.

In RDF 1.0, a reified statement is a member of Resources.:

"Reification of a triple {pred, sub, obj} of Statements is an element r of
Resources representing the reified triple ..."

The most straightforward thing to do is give it the capability to be an
***addressable*** Resource, ie, give it an ID (which would presumably be
relative to the serialized document).  This would maintain maximum
compatibility with current RDF.

But maybe contexts can do the job, or maybe we need both.  I'm unclear
whether identifying a statement with a context would make it an addressable
resource of type rdf:Statement.  What do you think?

Cheers,

Tom P

Received on Sunday, 14 October 2001 18:07:46 UTC