Re: IETF meeting agenda

Peter Deutsch (peterd@bunyip.com)
Mon, 19 Jun 1995 10:42:04 -0400


Message-Id: <9506191442.AA10177@expresso.bunyip.com>
From: Peter Deutsch <peterd@bunyip.com>
Date: Mon, 19 Jun 1995 10:42:04 -0400
In-Reply-To: Larry Masinter's message as of Jun 18, 10:39
To: masinter@parc.xerox.com, uri@bunyip.com
Subject: Re: IETF meeting agenda

g'day all,

Normally this would have come from Leslie, but she's out
of the office this week so I'm going to take a stab at it.


[ Larry Masinter wrote: ]
.  .  .
} I think this [URN proposals] and related topics should be our first agenda item at the
} next meeting. In our two hours, I would like to allocate at least 1
} hour to URN issues, and the remaining time between
} 
}  * URC issues
}  * revised charter
}  * URL issues (mailserver, z39.50?, revised URL standard)
} 
} Before submitting an agenda, are there any other topics that we should
} cover? 


Does it bother anyone else that with the closing of IIIR,
there is nowhere within the IETF that appears to be
considering the architectural issues implicit in all of
this acronym soup?

In this group we now have some maintenance being done to
work basically completed (ie. the proposed updates and
changes to URLs), we have some nitty gritty details of
proposals on URCs and URNs to debate, and we have an
upcoming discussion on the charter (which I suspect will
take the form of "Are we done? Is all the world a URL?")
but nowhere do I see serious time or consideration being
given to the fundamental question of whether our basic
model is necessary and/or sufficient for what needs to be
done _next_ year, rather than capable of being supported
in the next release of the WWW-based browsers.

There were draft proposals on architecture written up two
years ago but I still don't feel assured that they've
either been accepted by everyone or sufficiently debated
for alternatives to be refined. Beyond that, I am more
convinced than ever that we've focused far too much on
existing file-based metaphors ("URLs as file pointers and
URNs as simply pointers to URLs"). Basically, I think we
are lacking a number of elements needed for the next turn
of the technology crank (especially support for agents, be
they Knowbots, Hot Java or URAs but also for such things as
the KidCode discussion which is starting to take shape).

When my group proposed a new element in the puzzle (URAs),
the chair suggests that they don't even belong in this
working group. Well, if not here - where? And if not these
elements, what? And if there's no architecture behind it
all, how will we even tell if any new proposal warrants
keeping the group alive beyond the current mandate?

This may not be an issue to others, in which case we'll
pursue our work outside of this forum and bring it back
when we're done, but we'd much rather work with others to
develop a common infrastructure for what we see as the
missing elements. Right now, I'm not seeing the next
generation of technology taking shape here and that's
starting to bother me a bit.

Maybe we need more practical examples of what this means
for others to understand specific proposals (and those who
want to join our beta list are encouraged to contact Leslie
directly, as our new client contains support for things
which look remarkably like URLs, URNs, URCs _and_ URAs and
it's in the hand of beta testers now). Heck, maybe even
all the specific proposals we're going to see in the next
little while are wrong, but I think everyone should take a
few minutes to reflect on the process over the past year
or two and ask yourselves whether we've done much more
than refine existing work. Is that true? If so, is it
enough?

Personally I have the sneaking suspicion that, despite the
success of Mosaic, we're not done yet.  We've got browsing
under control (although there are things I still would
like which are possible) and we have some ideas about how
to do distributed indexing, but can anyone claim that
resource discovery (ie. the actual location of suitable
servers and the needed information) is a solved problem?
Can we even claim that our existing toolbox of techniques
even supports the necessary operations? Personally, I
think the answer is "no" and I don't think any of the
current URC/URN proposals will be enough, since they're
not actually addressing this problem.

With that in mind, perhaps we need some time set aside for
a "Whither Resource Discovery at the IETF?" discussion.
Maybe this falls under the "charter discussions" label, in
which case I know Leslie will want a timeslice in there to
outline our group's thinking on this and what the IETF can
do to help. And if this is out of order here, can someone
suggest where such discussions should go, other than the
corner of the bar with like-minded individuals? It's
currently the only viable alternative, which seems
suboptimal, to say the least...


					- peterd

--