W3C home > Mailing lists > Public > www-rdf-comments@w3.org > October to December 2002

RE: "Resource" (RDF vocabulary definitions)

From: Danny Ayers <danny666@virgilio.it>
Date: Wed, 27 Nov 2002 12:34:20 +0100
To: "Graham Klyne" <GK@ninebynine.org>, "Brian McBride" <bwm@hplb.hpl.hp.com>
Cc: "www-rdf-comments" <www-rdf-comments@w3.org>
Message-ID: <EBEPLGMHCDOJJJPCFHEFOEBLIKAA.danny666@virgilio.it>

what is the relationship *in the graph* (if
>any) between
>>>the resources referred to here? -
>>>
>>><rdf:Description rdf:about="http://example.org/something">
>>>and
>>><rdf:Description rdf:about="http://example.org/something#foo">
>>
>>None.  RDF treats URIrefs as opaque strings - those are different URIrefs
>>so as far as RDF is concerned they are different (rdf) resources.

I'm grateful for the clarification, and the pointer to Pat's comments [1] on
the topic. Until recently I was comfortable with URIrefs as names, but
trying to figure out how to use these artifacts of RDF in the real world has
given me something of a crisis of faith.

One the one hand there is the fairly intuitive but presumably incorrect
interpretation that if the resource being identified was an RDF document,
then statements made about that resource/document would apply to any RDF
(sub)graph represented by the document. On the other hand there is the
problem of how one graph can indeed be referred to in another. The idea that
a URI can be a name, locator or both is probably the root of my troubles,
when the name and the locator are one and the same thing it all seems
remarkably loosely defined.

In terms of the opacity, it rather feels like we start by talking about the
URIref in the same way we might with say the algebraic variables x, a, and
M. But then when we look at the resources we're talking about, we are asked
to consider this to be a person called Max.

The mapping appears to be that first the URIref is projected as a string,
then that string interpreted as a 'real world' URI. Irrespective of whether
there is anything at the end of the wire, this kind of coupling just seems
so loose as to be practically nonexistent. I can't help feeling that when we
treat the URIs within RDF syntax as pure names there isn't anything
associating the name with the named, and we end up being able to say nothing
about anything.

Out of curiosity, is there anything explicitly stated in the document suite
that would stop:

<rdf:Description rdf:about="http://example.org/aaa">

referring to the document at the URL

http://example.org/bbb

and

<rdf:Description rdf:about="http://example.org/bbb">

referring to the document at the URL

http://example.org/aaa

?

Anyhow, I am not a logician or semiotican and wouldn't be at all suprised
(or bothered!) to find that I'm talking from a fundamental misconception
here. The good thing is that I'm quite able to get on with coding (what I
hope will be) useful stuff. I have practical problems broadly related to the
point above in respect of containers, but if all else fails hackiness can
always come to the rescue ;-)

Cheeers,
Danny.


[1]
http://lists.w3.org/Archives/Public/www-rdf-comments/2002JanMar/0145.html
(Cut to the chase... search for '****')
Received on Wednesday, 27 November 2002 06:46:09 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 21 September 2012 14:16:31 GMT