To: sollins@lcs.mit.edu
In-Reply-To: sollins@lcs.mit.edu's message of Wed, 26 Oct 1994 17:10:11 -0700 <94Oct26.171018pdt.2762@golden.parc.xerox.com>
Subject: Re: Current URN syntax is unacceptable
From: Larry Masinter <masinter@parc.xerox.com>
Message-Id: <94Oct26.182231pdt.2760@golden.parc.xerox.com>
Date: Wed, 26 Oct 1994 18:22:28 PDT
I'll reply with some general remarks on managing URI@BUNYIP.COM
conversations, and specifically on the issue of case sensitivity in
URNs.
While I'm certainly willing to ask people to refrain from trying to
revisit issues that were already settled, I don't think this is the
chair's sole responsibility.
I think it is even more effective if those of you who have been
through the issues could courteously and privately respond if
newcomers bring up issues that we've effectively settled.
First, we should have some sympathy: it's sometimes hard to find the
discussion of old topics in the mail archives, especially given the
chaotic nature of the past discussions, and a little help goes a long
way.
Secondly, please respond *privately*, and see if you can work out the
issues offline. Sometimes, these discussions are merely one-way, but
often, they result in a deeper understanding. Often, battling old
battles on a public list results in 'flaming' that is more effectively
handled in a more responsible way in direct conversation.
================================================================
Re URN syntax:
We established after long discussion that we would like URNs to be
case insensitive, but we allow URLs to be case sensitive; the
reasoning was that although the transcribability requirement would
indicate a preference for case insensitivity, the requirement that
URLs also be used for describing existing Internet information access
protocols such as FTP, Gopher, and HTTP required case sensitivity.
I do think that further discussion of the syntax of URNs (whether they
have / or : and are case sensitive or insensitive) is
counterproductive at this time; we've spent most of our energy on it,
without any particular concern for the larger issues of the semantics
and mechanisms associated with them.
I think the most serious issue facing us is whether some of the
proposed URN systems are in fact more easily handled as URLs. For
example, the "XFN" system seems much more closely aligned with the
spirit and requirements for URLs than for URNs, just as the Mitra et
al proposal.
If we view these things as URLs, then (a) they need to follow the
guidelines for new URL schemes as far as trying to keep a common
syntax, and (b) they are allowed to be case sensitive. This might well
solve a large number of conceptual problems, although it still pushes
off the definition and application of a resource identification system
that is truly independent of the mechanism of resolution.
================================================================
On the general tasks before the uri@bunyip.com group:
We're working simultaneously on three issues: URLs, URNs, and URCs,
each of which may result in several documents (requirements,
standards, and auxiliary standards). Three of our documents are
(likely) to be promoted to RFCs soon, so that we can concentrate on
the others. I'll try to send out a revised 'status of the issues'
summary in the next few days. If those of you who are working on
various documents could let me know privately what your status is, I'd
appreciate it.