Re: Question for DNS propronents?

Larry Masinter (masinter@parc.xerox.com)
Wed, 14 Jun 1995 16:17:06 PDT


To: roxanab@attmail.com
Cc: uri@bunyip.com
In-Reply-To: roxanab@attmail.com's message of Wed, 14 Jun 1995 14:56:40 -0700 <95Jun14.145651pdt.2761@golden.parc.xerox.com>
Subject: Re: Question for DNS propronents?
From: Larry Masinter <masinter@parc.xerox.com>
Message-Id: <95Jun14.161718pdt.2761@golden.parc.xerox.com>
Date: Wed, 14 Jun 1995 16:17:06 PDT

> I don't know what the current status of the agenda is (do we still only
> have one session?), but given the number of different proposals on URNs 
> and the perception that the group needs to concentrate its efforts more, 
> it does not seem unreasonable to me to spend most of the session focusing
> on the various URN proposals and a plan for making some sort of progress
> in this area.

I view the meetings as an opportunity to air informal opinions and
have discussion. But most of the work and planning should happen on
the mailing list. I think we have an opportunity to make progress on
the URN proposal without interfering with the progress on URCs and we
should do so.

> Personally, I do not think we will come up with a single solution that 
> combines all the proposals and I think even if we did, such a solution would
> probably be worse than any one of the given proposals. So what I would 
> suggest is that perhaps instead of trying to focus on converging the
> solutions maybe we focus on interoperability for now (even if it's 
> only among some of the proposals and not all) at the syntax and protocol
> level. Based on my reading of the proposals I think this doable. What do
> other folks (especially those with proposals) think about this?

I've been hoping for 'selection' rather than 'convergence', i.e., that
we would discard some of the proposals as 'unworkable' and merge the
rest. As I mentioned before, I think we can separate out

   syntax, structure and semantics of names
   syntax, structure and semantics of descriptions (characteristics)

from

   mechanisms for relating names, descriptions, and resources.

The current charter of the URI working group is to define the syntax
structure and semantics of resource identifiers, by the way, and _not_
to establish standards for protocols. The only purpose for
implementation experimentation is as an existance proof that, for
example, the name syntax does not hinder the ability to build an
effective resolution mechanism.

I think in the end we will be faced with choices where there is no
good reason to choose one form over another (what is the delimiter
between naming authority and name) and that these will be the hardest
for us to reach because most of the opinions are informed by instincts
that are hard to express (familiarity).

If my posts are a little disjointed the last few days, it's because
I'm using an awkward mechanism for dealing with my mail, and I
apologize.

I would like to encourage anyone who has a personal scorecard on 'URN
proposals and my opinions on them' to post it.