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

RE: "Resource" (RDF vocabulary definitions)

From: Brian McBride <bwm@hplb.hpl.hp.com>
Date: Mon, 25 Nov 2002 14:32:38 +0000
Message-Id: <5.1.0.14.0.20021125141805.03164fa0@0-mail-1.hpl.hp.com>
To: "Chris Catton" <chris.catton@btopenworld.com>, "Dave Beckett" <dave.beckett@bristol.ac.uk>
Cc: "seth" <seth@robustai.net>, "www-rdf-comments" <www-rdf-comments@w3.org>

At 13:17 25/11/2002 +0000, Chris Catton wrote:
>(This thread began in rdf interest)
>
>OK, my understanding is deepening I think, but I also think there is still
>be a problem here.

You may well be right.  Lets go through the logic of it and see.  However, 
I think you are into an area that properly belongs in web architecture, not 
just RDF.  What exactly does a URI, or URL for that matter identify.  We 
architecture doesn't seem to be that well defined on this point, and it is 
all a bit tricky.


>In my *rdf* document I write
><http://example.org/somepage#MotorVehicle>  rdf:type
><http://example.org#HtmlFragment>
>
>In my *html* document at http://example.org/somepage, i can write (stupidly
>perhaps, but perfectly legally)

It is not illegal to write stupid things, just unwise.  RDF has no 
pretensions to stop people writing things that are wrong or stupid.  I've 
no doubt one can construct cases that don't make sense, e.g.

   <rdf:Description>
     <rdf:type rdf:resource="&eg;sheep"/>
     <rdf:type rdf:resource="&eg;tulip"/>
     ...


><a name="MotorVehicle">model t ford type thing</a>
>
>and in the header on the same page (as suggested at
>http://www.w3.org/RDF/FAQ#How )
>
><rdf:Description rdf:ID="MotorVehicle">
>   <rdf:type rdf:resource="http://www.w3.org/2000/01/rdf-schema#Class"/>
>   <rdfs:subClassOf
>rdf:resource="http://www.w3.org/2000/01/rdf-schema#Resource"/>
></rdf:Description>
>
>Now:
>1.      My html document is referred to from an rdf document and so 
>treated as an
>rdf document. (stated in http://www.w3.org/TR/rdf-concepts/#xtocid103660 )

Err, where exactly was your html *document* referred to from your rdf 
document?  Do you mean:

   <http://example.org/somepage#MotorVehicle>  rdf:type ...

This does not refer to your html document.  It refers to the rdf resource 
identified by:

    http://example.org/somepage#MotorVehicle

The HTML document is a *representation* of a resource, as best I can 
understand the current state of web architecture.  There could be another 
representation of that URI, e.g. in PDF format.



>2.      As an rdf document a reference to
>http://example.org/somepage#MotorVehicle refers to whatever the owner of
>http://example.org says it is - her car perhaps.
>3.      Her car is not an HtmlFragment - so we have a category error.
>
>So I don't see how I can reliably reference an html fragment from an rdf
>document.

Well, I guess you'd have to define the URI:

   http://example.org/somepage#MotorVehicle

to mean the HTML fragment, in contradiction to your statement above.

>Common sense might tell us what was meant, but I don't see the
>rules that let a machine decide.

An RDF processor never really 'knows' exactly what is meant by a URI.  It 
doesn't have to.

>Anyone see a way out of this?

What you need is a URI that you know refers to the html representation of 
the resource.  I guess (YUK) we could define a urn or uri scheme for this:

    rep:text/html:http://example.rog/somepage#MotorVehicle

which refers to a mime-type specific representation of URI.

>Paul
>Prescods post Re: a URI is a name (tel uri scheme and VCARD RDF) on
>rdf-interest suggests we don't need 2 URIs because we can use rdf to
>distinguish the meaning.  If we cannot actually reference the fragment
>reliably, this is not true.

I haven't read that - I better go look.

Brian
Received on Monday, 25 November 2002 09:31:27 GMT

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