RE: TB16 Re: Comments on arch doc draft

> -----Original Message-----
> From: Michael Mealling [mailto:michael@neonym.net]
> Sent: 01 July 2002 23:14
> To: Tim Berners-Lee
> Cc: Dare Obasanjo; Ian B. Jacobs; Tim Bray; www-tag@w3.org
> Subject: Re: TB16 Re: Comments on arch doc draft
> 
> 
> 
> On Mon, Jul 01, 2002 at 05:51:30PM -0400, Tim Berners-Lee wrote:
> > Yes. URNs are not dereferencable and therefore should not be used for
> > namespace names.  It deprives someone who does not know the name
> > of the ability to look it up and get useful information about it.
> > It especially deprives a machine of that possibility.
> 
> Tim, 
> 
>   Either the web architecture is URI scheme agnostic or it isn't. If
> the TAG is coming up with architecture that is scheme dependent then
> IMNSHO, its broken.

Yes... I too would like to see us define as much as we can in a way that is
URI scheme agnostic. 

We seem to be struggling with at least two different world views about the
things identified by URI references with fragment IDs and the things
identified 'simply' by URI, and whether the URI scheme used affects the
nature of the resource being identified eg. assertions of the style "http
URI identify documents."

World View A
-------------
a) A URI (without fragment ID) identifies a resource.

b) "A resource can be anything that has identity." [RFC2396] presumably
identifiable by URI.

c) "Not all resources are "retrievable"; e.g. human beings, corporations,
and bound books in a library can also be considered resources." [RFC2396]

d) A resource is a time varying conceptual mapping to a set of equivalent
values which are either resource identifiers or resource representations.
(paraphrasing [Fielding]).

e) A fragment ID in a URI Reference is interpreted only after the retrieval
of a representation of the resource referenced by the URI entailed in the
URI Reference. "The format and interpretation of fragment identifiers is
dependent upon the media-type [RFC2046] of the retrieval result." [RFC2396].

f) "A fragment identifier is only meaningful when a URI reference is
intended for retrieval and the result of that retrieval is a document for
which the identified fragment is consistently defined." [RFC2396] (using
document synonymously with (resource) representation).

World View B
------------
(which seems to specialise around HTTP scheme URI)
a) A URI Reference (with or without a fragment ID) identifies a resource.

b) "A resource can be anything with identity." [2396] presumably
identifiable by URI reference.

c) "Not all resources are "retrievable"; e.g. human beings, corporations,
and bound books in a library can also be considered resources." [RFC2396]

d) Resources identified by URI (without fragment IDs) SHOULD have network
retrievable representations (view might be restricted to HTTP scheme URI).

e) Resources that do not have network retrievable representations are
(SHOULD BE) identified by URI references that include a fragment ID (some
argue that this is an RDF convention - and again may be a view restricted
just to HTTP scheme URI).

f) "HTTP scheme URI identify documents." (here I think document is an
abstract work and what is retrieved is a representation of a document).

The principle difference between these world views (IMO) is whether two URI
references that differ only in their fragment ID are regarded as identifying
the same (WV-A) or different resources (WV-B), or stated differently,
whether a fragment ID is part of a URI (RFC2396 states clearly - not).

WV-B (at least for HTTP scheme URI) seems to distinguish between URI
identifiable resources (which (should?) have retrievable representations)
and URI+fragment identifiable resources which can include abstract concepts
such as an ontological notion of a Person (eg.
http://www.ksl.stanford.edu/projects/DAML/ksl-daml-desc.daml#Person) as well
as real-world objects without network retrieveable representations.

Refinement to either of these prototypical world views welcome, (or indeed
further distinct views).

<snip/>

> 
> -MM

<snip/>

> -- 
> --------------------------------------------------------------
> ------------------
> Michael Mealling	|      Vote Libertarian!       | urn:pin:1
> michael@neonym.net      |                              | 
http://www.neonym.net

Regards

Stuart Williams
--
[Fielding]
http://www.cs.virginia.edu/~cs650/assignments/papers/p407-fielding.pdf

Received on Wednesday, 3 July 2002 10:08:31 UTC