Re: URI etymology

Sean,

I admit I may be wrong about the fact that http: URIs identify
documents.
As a matter of fact, it just occured to me that an HTTP server could
return me a "204 No content" code, which is a success (hence I know the
URI is "valid"), provide meta-data about the resource in the HTTP
header, hence the resource may be something else than a document.
However...

On 11 Jun 2001 16:21:09 +0100, Sean B. Palmer wrote:
> What about http://example.com/# - is that a generic document? What
> about http://purl.org/dc/elements/1.1/title - is that a generic
> document? I suggest that you dereference it, and once and for all blow
> apart your notion that anything on an HTTP server is a document.

The fact that they do not resolve does not mean, for me, that they are
no document !
It may that the server is unreachable or to busy... how would I know ??
Answer: 204 code ! Which I don't think purl.org returns.

> You just ask for a URL, you don't tell the server what should be there.

I don't tell ! I only guessed the "http:"prefix tells me... But you got
it, I changed my mind :)

> Onto cases where URLs represent both a "namespace" and also return a
> document. That document is simply a representation of a resource, not
> the resource itself, so is there really any problem there?

There is to me.
I would say that a URI may resolve to different *instances* of the same
resource.
e.g. differing by their encoding format, language, version, etc...
I don't take a description of a thing to be an instance of that thing
(document about a namespace, picture of a person, map of a city)
this is *not* a matter of encoding format IMHO.

But OK, would http://www.w3.org/1999/xlink/ return nothing with a 2xx
code, I would admit it identifies the namespace unambiguously -- mark my
words ;-)

  Pierre-Antoine

Received on Monday, 11 June 2001 12:03:44 UTC