DRAFT Minutes, URL BOF

Larry Masinter (masinter@parc.xerox.com)
Sat, 14 Dec 1996 15:11:51 PST


To: uri@bunyip.com
Subject: DRAFT Minutes, URL BOF
From: Larry Masinter <masinter@parc.xerox.com>
Message-Id: <96Dec14.151151pst."2695"@golden.parc.xerox.com>
Date: Sat, 14 Dec 1996 15:11:51 PST

Minutes by Andy Mutz & Ryan Moats, merged and edited by Larry Masinter

Please send me corrections within the next few days; I'll then submit
to minutes@ietf.org

================================================================
The URL BOF was convened to decide if there was a need to form a URL
working group for either of two purposes:

- Advance URL syntax draft to Draft Standard
- Create an RFC for BCP to describe the process for creating new URL
  schemes


There is a new draft-fielding-url-syntax-NN.txt (-00 in internet
drafts, -01 and -02 submitted) which combines the generic syntax from
RFC 1738 and 1808.

In addition, there are at least half a dozen new URL schemes being
proposed. Do they match the generic syntax in
draft-fielding-url-syntax?  Some of the individual schemes that were
in 1738 have been split out into separate drafts. Should a working
group review these?  RFC 1738 states a process for registering new URL
schemes will be developed.  This has never been documented and needs
to be.  Further, it is becoming clear that URL schemes constitute a
namespace, so how does registration occur?

Many people expressed the feeling that no new working group is needed
to process draft-fielding-url-syntax (except via mail on the mailing
list), but a working group is needed to deal with the process. Harald
Avelstraad had a web page about _his_ requirements for URL schemes,
(http://www.apps.ietf.org/apps) and this was proposed as a starting
point.

There is belief that IANA has a file listing registered URL schemes.

Right now, URL schemes need a standards track document.  That's ok for
the most part, but private use URL schemes are around (e.g., the
"about" scheme in Netscape Navigator and the "aol" scheme in the AOL
web interface.)  Some browsers allow plugins to recognize new URL
schemes.  Is a namespace required for definition to avoid collision?

(Harald) "What the IETF thinks about URL schemes is somewhat irrelevant
in the marketplace. There's a java: scheme and a clsid: scheme in use.
What I would like to see is 'This is community consensus on how to
handle new URL schemes' Therefore I want to see community input into
this, so I don't want to write this.  The best way to achieve
consensus is to have a classic WG, with a chair that believes we can
get there, membership from the big URL players (Microsoft, Netscape,
Web Consortium, etc.) so that the WG will be 'listened to'.  If we
find a chair and distinguished members, then we should have a WG.
Otherwise, we continue on the current path of requiring a standard
track document for each URL scheme."

We need some general advice on URL schemes; for example, some
proposals use "//" as 'top level' even though the syntax after the
'//' is not an internet host name. There are some proposed URL schemes
where the URL doesn't really locate a resource on the Internet, e.g.,
"phone", "tv", "fax".  These kinds of design questions should be
answered consistently.

"Given that there are so many URL schemes backlogged as I-D, shouldn't
there be a single WG to handle this?"  The point of listing the
Internet Drafts wasn't that the WG should review schemes in
perpetuity, it was to show the problem and that a procedure was
needed.

(Harald) "The WG will develop and apply a process, and recommend a
process as a result, but WG's are not permanent and the URL working
group won't be either."

The working group will NOT consider the generic syntax draft. There's
some desire to make sure that the Draft Standard gets advanced before
the working group is chartered, just to keep the boundary clear.

Some of the folks from web browser makers felt there was real value in
having the accepted IETF process documented.  At the meeting,
people from W3C and various web browser companies endorsed the
formation of the working group and pledged to participate.

- editors: Larry Masinter volunteered to co-edit the vetting and
   advice document. 

- mailing list
  The uri@bunyip.com list would be used to announce the formation of
  the working group, but a new mailing list would be set up with new
  subscriptions. The 'uri@bunyip.com' list is saved for discussing
  generic URI issues that are not otherwise being handled by another
  group.

  To subscribe to uri@bunyip.com, send mail to 
  uri-request@bunyip.com.

 
- chair: Rich Petke volunteers to co-chair a new WG.

- charter (To devise the process for vetting new URL schemes
   and give advice to authors, history.)

Milestones to be defined, but should not be a long-term process.
This work should be short-term; ideally a few months work.

A number of other namespace vetting schemes exists now.  One for MIME
types.  One for SMTP extensions.  Review of these schemes would be
useful.