From: jmf@kudzu.cnidr.org (Jim Fullton)
Message-Id: <9308101102.AA23570@kudzu.cnidr.org>
Subject: Minutes - Again
To: uri@bunyip.com
Date: Tue, 10 Aug 93 7:02:10 EDT
Folks - I sent these in some time ago (at or near the deadline) and
*thought* I posted them to the list. Some people got them, but quite a
few seem to have not. My apologies if this is your second (or third)
copy.......
Jim
============================================
Jims comment with the original message.....
I tried to summarize what we actually did, rather than what we
argued about.
Jim
Minutes - URI Working Group
Co-Chairs - Alan Emtage, Bunyip Information Systems
Jim Fullton, CNIDR
Session I
After introductions, Tim Berners-Lee presented the current URL draft.
After discussion, the following changes were proposed:
1. Remove the AFS type.
2. Add the "mailto" type - mailto:timbl@cern.ch - designed to
be a generalized mail transmission URL. Designed to help solve the
problem of how to contact an author
3. Wrapping - the formal URL wrapper is defined as <>
A significant portion of Session I was spent discussing the merits of
creating a more generalized URL specification, with the current spec as
a specific case. Peter Deutsch advocates the creation of this far more
generalized specification and will examine the issue further before
the next IETF. During these discussions, it became clear that the
allowable character set needed refining.
The three general proposals were:
-no restrictions on character sets
-significant restrictions on the character set; i.e. the current
document
-looser restrictions on the character set, but stay with the
current proposal
Further discussion of URL issues was put off until Session III.
Chris Weider closed out Session I with a tutorial on current URN issues.
Session II
Session II was spent on URN issues. Karen Sollins made a presentation
describing a system under development at MIT - her slides are attached
to the hard-copy minutes.
URN's were discussed further in session II. The group agreed that
several issues needed to be resolved before URN's could be properly specified:
-detection of duplicates
-fragment specifiers
-version specifiers
-type specifiers
After vast amounts of discussion, fragment, version, and typing
information was moved to a Uniform Resource Citation object, which
has not received a formal name. The URN was defined as a URN
identifier (RURNS), a Naming Authority string and an opaque string defined
by the naming authority as being unique within its name space. The whole
URN is wrapped in <>, just like a URL:
<URN:NA:OpStr>
A URN should occupy the same character space as a URL, for ease of
applications upgrades.
Session III
The discussion of URLs restarted in Session III. Mitra led a discussion
about modifications to the URL character set, and the group agreed to modify
the URL draft to allow the use of the following characters:
: ; = ' "
Other characters not included in the draft are excluded.
A vote on the modified draft was held, with group recommending adoption
of the modified draft by a vote of 29 to 1.
The URN format as described in Session II was informally agreed upon,
which led into a discussion of naming authorities. It is assumed that
many naming authorities will exist, including authorities within an
organization whose existence will not be known outside the organization.
Next Time
At the next IETF, and agenda item will be specified for a "sub-ID"
discussion to define the format of the naming authority string and
opaque string.
Jim Fullton