W3C home > Mailing lists > Public > www-rdf-interest@w3.org > January 2003

What does a URL identify: pragmatic approach

From: Brian McBride <bwm@hplb.hpl.hp.com>
Date: Tue, 28 Jan 2003 19:44:28 +0000
Message-Id: <5.1.0.14.0.20030128191858.053aeac0@localhost>
To: www-rdf-interest@w3.org

On the tag list there has been a long thread on naming and what URI 
references identify.  I don't want to fan the flames on that thread or 
burden the tag members with deleting more unwanted mail, but I did want to 
record an approach I'm trying out.  I hope this doesn't generate too much 
traffic here.

If you point your browser at:

   http://www.w3.org/2001/sw/RDFCore/20030123-issues/

you should see the RDFCore's last call issues document.

When you do a GET on that URI, the server returns a representation of:

   http://www.w3.org/2001/sw/RDFCore/20030123-issues/Overview.html

because that is the default document for a directory.

So my conceptual model is this:

http://www.w3.org/2001/sw/RDFCore/20030123-issues/ identifies rdfcore's 
last call issues list.

http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-01 identifies a 
particular issue in that list

http://www.w3.org/2001/sw/RDFCore/20030123-issues/Overview.html identifies 
an xhtml document, a representation of which is returned when I do a GET on 
the issues list.  This is a document which provides useful information 
about that list which the server is configured to return when a GET on the 
abstract concept is received.

http://www.w3.org/2001/sw/RDFCore/20030123-issues/Overview.html#pfps-01 
identifies an element in that xhtml document as defined by the appropriate 
mime-type.

http://www.w3.org/2001/sw/RDFCore/20030123-issues identifies the directory 
which contains Overview.html.

So pragmatically speaking, I seem to have different names for the different 
things I want to name, and it just seems to hang together.
I can make rdf statements about the issues list, individual issues, the 
document and elements of the document and its clear which is which.

Pragmatically speaking, is this a good model to follow?

And architecturaly, I'm left wondering whether there is a useful 
generalization built around the fact that many web servers allow GET's on 
one URL to be mapped to a different URL.  What I have done here is to use 
such a mapping to have the web server map from a concept (a last call 
issue) to a document giving information about that concept.

Brian
Received on Tuesday, 28 January 2003 14:45:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:57 GMT