On Tue, 21 Sep 2004 08:02:48 +0100, Pete Johnston <[log in to unmask]>
wrote:

>If I want to make my identifier "resolvable", the HTTP protocol, the DNS
>system and the ubiquity of desktop tools gives me a means of allowing anyone
>to dereference my identifier to obtain a representation of my resource. As
>the owner of the URI I can choose to provide a representation or not. I
>might provide different representations to different users or to provide
>different representations depending on how a user makes their request
>(through e.g. HTTP content negotaiation).

And (prompted by Patrick Stickler's wise words over on www-rdf-interest [1])
if I want to talk about my representations separately from the resource, I
just make sure I use different URIs to identify those representations:

<http://myorg.org/collection/6789> a <http://purl.org/dc/dcmitype/Collection> .

<http://myorg.org/collection/6789/intro.html> a
<http://purl.org/dc/dcmitype/Text> .

<http://myorg.org/collection/6789/intro.jpeg> a
<http://purl.org/dc/dcmitype/Image> .

<http://myorg.org/collection/6789/intro.mp3> a
<http://purl.org/dc/dcmitype/Sound> .

etc etc

And all of those representations might have a different creator, who, in any
one case, is almost certainly a different agent from the creator from the
collection.

If you type http://myorg.org/collection/6789 into the address bar of a Web
browser, your browser may display an HTML document or a JPEG image or
download an MP3 file. You may get a 404.

But you should not assume from that behaviour that it follows that
http://myorg.org/collection/6789 denotes an HTML document or a JPEG document
or an MP3 file or nothing. After all, I might choose to deliver a different
representation every day of the week: that's perfectly OK and doesn't in any
way change the usefulness of http://my.org/collection/6789 as a persistent
identifier.

It is up to the owner of the URI http://myorg.org/collection/6789 to say
what that URI denotes, and if the owner says quite unambiguously that it
denotes one thing and it is going to go on denoting that one thing for ever,
then there is no room for ambiguity - and it's not "socially responsible"
(to use Patrick's term) for someone else to come along and say it denotes
something else.

Pete

[1] http://lists.w3.org/Archives/Public/www-rdf-interest/2004Sep/0145.html