To: Michael.Mealling@oit.gatech.edu
In-Reply-To: Michael.Mealling@oit.gatech.edu's message of Tue, 20 Sep 1994 15:33:41 -0700 <94Sep20.153357pdt.2760@golden.parc.xerox.com>
Subject: Re: Meta info tags
From: Larry Masinter <masinter@parc.xerox.com>
Message-Id: <94Sep20.171150pdt.2760@golden.parc.xerox.com>
Date: Tue, 20 Sep 1994 17:11:44 PDT
Someone (me?) asked:
> > What's the intent for the relationship between URCs and HTML headers?
and I commented:
> I think the requirement section of draft-ietf-uri-urc-spec-00.txt
> needs to be elaborated in order to answer the question.
And Michael asked:
> Is it appropriate that implementation issues like that go into the
> requirments document? I would think that would be an implementation
> issue. I may be wrong or atleast splitting very fine hairs....
Sorry for the misunderstanding, I don't think the requirements section
of the URC document needs to specify implementation details. I asked
what the *intent* of URCs are: in particular, are they intended to be
part of the HTTP protocol or any other protocol; useful for the
negotiation between client and server over whether a particular
document is suitable for display or retrieval, caching, useful in
Z39.50.
Currently, the URC requirements section doesn't give much of a clue as
to the intended use. I'll admit that the URL and URN requirement
documents don't contain a lot of contextual use information either,
but that's because we *knew* applications for URLs (and, in future,
for URNs), but the application scenarios of URCs are missing; there
are not even any pointers or references.
I asked
> I'm wondering, by the way, whether there's a different representation
> for URC-as-specification ("a set of attributes that you might expect
> to find") vs URC-as-description ("what are the characteristics of this
> here document").
and Michael replied:
> Can you ellaborate? I would expecta heck of a lot of things that
> would/could be in both depending on what function the user is
> currently doing....
Of course there are a lot of things that are in both. However, in one
case, you're describing a set of critera, while in another, you're
describing a set of open-ended bits of metainformation. Think of the
difference between the media type given in Accept: and what is
specified when you return a Content-Type:. They're both talking about
the format of the document, but one indicates what you want, and the
other indicates what you got. Are there other kinds of correspondences
with other kinds of meta-information?