Re: Re: Meta info tags

Larry Masinter (masinter@parc.xerox.com)
Fri, 23 Sep 1994 00:19:50 PDT

To: mitra@path.net
In-Reply-To: mitra@path.net's message of Thu, 22 Sep 1994 22:16:33 -0700 <94Sep22.221645pdt.2763@golden.parc.xerox.com>
Subject: Re: Re: Meta info tags
From: Larry Masinter <masinter@parc.xerox.com>
Message-Id: <94Sep23.001954pdt.2762@golden.parc.xerox.com>
Date: Fri, 23 Sep 1994 00:19:50 PDT

I said:

>> but the application scenarios of URCs are missing; there
>>are not even any pointers or references.

and Mitra replied:
> Larry - I think the application scenarios are all over the place, anywhere
> we have meta-information being passed around, for example Gopher+, HTTP
> "Head" or Prospero. What we dont have is a common way to pass meta info
> between these protocols.

I'm sorry if I wasn't clear. What I meant to say was that the
application scenarios are missing FROM THE DOCUMENT. I know they're
all around us. However, I believe, from my observation of the
discussion around URCs, that much of the confusion and difficulty in
resolving URC issues is because not everyone has the same context, is
aware of all of the application scenarios that others are thinking
about.

I think this was a problem with the URL discussion: some folks were
primarily thinking about using URLs primarily as they exist in World
Wide Web, and primarily for accessing material that can be cached,
while others were considering URLs in Archie, SOLO, etc., in
situations where relative URLs weren't important, etc. I don't want to
go back and revisit the issue for URLs, but I'm hoping we'll make more
progress on URCs if we can put at least REFERENCES to some of the
application scenarios in the internet draft:
Use in Gopher+
Use in Prospero
Use in HTTP HEAD
Use in Z39.50 ?
Use in Harvest?
Use as in RFC 1357 ?

It doesn't have to be exhaustive, but I'd like to at least cover what
the members of this working group have in mind.

> Back to the original question - I think specifications of a minimum set of
> attributes belong with particular programs, or in applicability documents.
> e.g.
> "When you see a request for the HEAD of a document return a URC with at
> least the Author, Date and URL fields"

Do you think each individual program gets to decide what the 'Author'
field means? If you're going to share URCs from one application to the
next, is there some least-common set that must be present?