W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > October to December 1997

Re: Removing External References

From: Ellis Cohen <ellis@crystaliz.com>
Date: Thu, 18 Dec 1997 14:26:45 -0500
Message-ID: <349978F5.2F1F@crystaliz.com>
To: Jim Davis <jdavis@parc.xerox.com>
CC: w3c-dist-auth@w3.org
Jim Davis wrote:

> Speaking for myself only, I don't agree.  I don't see the need for naming
> or numbering external members (as opposed to ordering *all* members, which
> I have argued for)

Perhaps there's a misunderstanding here.  I was suggesting naming and
numbering *all* members.  Both ADDREF and PUT take a Position header,
much like your ordering proposal.

The difference is that, for numbering, the Position header can take an
index as an alternative to a resource URI, which allows multiple
instances
of the same external resource at different places.

In talking to Judith Slein at the IETF meeting, she mentioned that this
was an
issue for Xerox as well, since a compound document might indeed need to
have
references to the same external document in multiple places.

> and I also believe that external members are quite
> useful even without any naming or numbering at all.  As evidence for the
> latter, there is a Xerox product called DocuShare which is a Web-based
> repository for documents.  In the DocuShare model, a "collection" (they
> don't call them collections but that's okay) can store both external
> members (URLs) and internal members.  There is no ordering or numbering.
> (Everything has a name, it's true, but DAV already provides a name in the
> form of the displayname property.)
> 
> So I must respectfully deny your assertion.  We need external members, and
> we don't need numbering.  Removing external members would hurt, and I see
> nothing to gain by holding out for "numbered" external members.

There might be cases like DocuShare in which unnamed/unnumbered
external members are useful, but I believe such cases represent a
relatively
small minority of the uses of external references in collection-like
entities
now and in the future.

But, given that you're a true believer in unnamed/unnumbered external
references,
and given where we are in the evolution of the spec, I'll retract my
proposal
to remove external references from the spec, and instead integrate
unnamed/unnumbered
external references with my proposal.

  -- Ellis
Received on Thursday, 18 December 1997 14:21:09 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:43:44 GMT