- From: Dan Brickley <danbri@w3.org>
- Date: Thu, 7 Sep 2000 10:45:43 -0400 (EDT)
- To: Dan Connolly <connolly@w3.org>
- cc: Miles Sabin <msabin@cromwellmedia.co.uk>, xml-uri@w3.org
On Thu, 7 Sep 2000, Dan Connolly wrote: > Miles Sabin wrote: > > > > Simon St.Laurent wrote, > > > I think the 'rest of us' might well benefit from clearer > > > distinctions between URIs and the resources they identify, > > > > But Simon, that's a pretty tall order, because there's no > > end to the number of distinct resources a URL (nb. UR*L*) might > > locate. > > Not so: every URL is a URI[1]; every URI identifies exactly > one resource[2]; hence every URL identifies exactly one resource. > > [1] The > term "Uniform Resource Locator" (URL) refers to the subset of URI > that identify resources via a representation of their primary access > mechanism (e.g., their network "location"), rather than identifying > the resource by name or by some other attribute(s) of that resource. > > [2] "An identifier is an object that can act as a reference to > something that has identity." > > http://www.ietf.org/rfc/rfc2396.txt The text you cite, [2], leaves open the question of whether the 'something' that is referenced may differ over time. At any tick of the clock, the identifier 'can' act as a reference to some thing; as time passes and the world changes, it might subsequently fail to refer, or pick out a different thing-with-identity. This seems to be one of the various factors contributing to confusion; while "every URL identifies exactly one resource" might be true at any point in time, when people aggregate metadata about these URIs/resources that spans several years, it isn't clear what the story is. > > John Cowans example (I forget the URL, it was a time- > > server of some sort) was pretty good. Does that URL identify > > entities (particular sequences of bits on any given occasion), > > no; those entities represent the state of the resource > identified. > > > transient HTML documents (again changing over time); > > no. I believe there's a good case to be made for modeling these entities as resources too. They're *not* the resource named by the URI, but something else related to the named resource. We could usefully name the relationship that holds between these entities and the more abstract resources that they're manifestations/rendings of. This way we could talk about them using any frameworks for resource description we happen to have lying around... Dan
Received on Thursday, 7 September 2000 10:45:49 UTC