W3C/IETF coordination meeting, June 2002

This is a record of an IETF/W3C coordination teleconference,
Thursday, June 27, 2002.

Sorry it took me so long to get this together.

The > quoted bits are from the agenda.


Attendance:

  - Tim Berners-Lee, W3C
  - Scott Bradner, IESG
  - Dan Connolly, W3C, (chair/scribe)
  - Leslie Daigle, IAB
  - Ned Freed, IESG
  - Hugo Haas, W3C (for part of the meeting)
  - Ian Jacobs, W3C communications
  - Michael Mealling
  - Joseph Reagle, W3C (XML Signature WG chair)
  - Ralph Swick, W3C

regrets:

   - Larry Masinter
   - Yves Lafon, W3C
   - Alison Mankin, IESG


== Next Meeting

RESOLVED: to meet again
	Tuesday, November 12, 2002, at 12:00:00 PM Boston time
	Tuesday, November 12, 2002, at 17:00:00Z
http://www.timeanddate.com/worldclock/fixedtime.html?day=12&month=11&year=2002&hour=12&min=0&sec=0&p1=43


>   -- URI spec update
> 
> # update RFC 2396 Larry Masinter (Wed, May 01 2002)
> http://lists.w3.org/Archives/Public/uri/2002May/thread.html#0
> 
> also: digression in
> 
>   TAG minutes 24 Jun
>   http://www.w3.org/2002/06/24-tag-summary


Several participants observed that there are a lot of
parties interested in any change to the URI specification;
large/costly discussions are quite likely.

Leslie Daigle asked if the W3C URI group had discussed
this revision? Dan Connolly and Michael Mealling noted that the W3C
URI group was in a slow transition between being
a W3C Interest Group and a W3C Coordination Group,
but that no, it hadn't discussed this update.

Mealling and Ned Freed noted that the proposed
changes were mostly minor (IPV6 syntax, etc.)
except for the integration of Internationalized URIs.

Connolly reported that he's undecided, after considerable
deliberation, on whether the URI RFC should be updated to
include "IRI"s at this point.

Freed suggested that an IETF working group is perhaps in order;
the process of assembling a charter is the traditional way
to scope potentially costly discussions.

ACTION Freed: to contact Larry Masinter
about drafting a charter for a revision to RFC2396.

no constraint on whether I18N stuff is in scope or not.

[postscript: this has been done. See:
  # proposed charter items for possible URI working group
  Larry Masinter (Sat, Jul 20 2002)
  http://lists.w3.org/Archives/Public/uri/2002Jul/0016.html
]


>   -- FYI: public-ietf-w3c announced (5Jun)
>  
http://lists.w3.org/Archives/Public/public-ietf-w3c/2002Jun/0001.html

Noted.


>   -- W3C WGs registering media types:
>         OK to point from IANA registry to W3C spec,
>         in stead of an RFC?
> 
>   cf
> 
>    TAG Finding: Internet Media Type registration, consistency of use
>    http://www.w3.org/2001/tag/2002/0129-mime


Freed summarized the issue as a request to relax a constraint
currently specified in an IETF BCP:

  "Proposals for media types registered in the IETF tree must
  be published as RFCs."
   -- 2.2.8.  Publication Requirements
	Multipurpose Internet Mail Extensions (MIME) Part Four:
                        Registration Procedures
	http://www.ietf.org/rfc/rfc2048

Ian Jacobs summarized interaction between recent W3C TAG
findings
  TAG Finding: Internet Media Type registration,
	consistency of use
  http://www.w3.org/2001/tag/2002/0129-mime
and end-game work in the W3C XML Protocol working group.

Tim Berners-Lee elaborated:  W3C specs have specified a language...
language specifications are connected to Web Architecture
via the MIME type registry... sometimes the material submitted
to the MIME registry has said more than the W3C REC-track
spec says... The questions in the MIME registry are good ones;
the answers belong in the spec... so the W3C TAG decided the
answers to the MIME registry questions should be in the W3C spec...

Freed described an earlier proposal that should perhaps be revisited:
a std. facet (ala the prs. and vnd. facets); registration of
media types (e.g. application/std.w3c.foo) could be delegated
to standards organizations such as W3C. He estimated the cost
of updating RFC2048 to be fairly low, though very near-term
progress would compete with the upcoming IETF meeting in
Yokohama (14-19July); e.g. the Internet Draft publication process
is on hold until after that meeting.

Joseph Reagle reacted positively to this proposal, as a long-term
solution, but noted that several W3C specifications are in
the process of registering media types currently.

Berners-Lee suggested using the existing DNS and HTTP mechanisms
as a long-term strategy for delegation of parts of the registry.
The std. facet procedure involves IANA for no good reason.

Freed and Michael Mealling noted that this sort of thing
has been proposed before, with vocal opposition. The opposition
finds management of the registry very valuable, is satisfied
with the way IANA manages it, and is unwilling to risk
delegating it to other parties.

Berners-Lee said it's OK for IANA to manage publication of
the IETF process, but to suggest that peer organizations
cannot do likewise is counter to Web Architecture.

[... the scribe didn't follow the technical discussion
that followed... something about using URIs
in HTTP Content-Type header fields ...]

Joseph Reagle recalled, from our 11 March meeting,
support for the idea of having URIs for media types,
and asked about any relevant progress.

Daigle reported that the documents specifying URNs
for media types had recently been approved by the IESG;
they're awaiting publication by the RFC editor.
this means that URNs can go live after IAB/IESG coordination.

Joseph Reagle asked about the status of a specific pending request:

  From: Joseph Reagle <reagle@w3.org>
  To: internet-drafts@ietf.org, ietf-types@iana.org
  Date: Wed, 5 Jun 2002 17:21:55 -0400
  Cc: ietf-xml-mime@imc.org, xml-encryption@w3.org
  Message-Id: <20020605212156.7576487E@policy.w3.org>
  Subject: Request to publish ietf-draft of
	XML Encryption media-type registration
http://lists.w3.org/Archives/Public/xml-encryption/2002Jun/0011.html19:33:44 
Freed clarified: that's fine for getting an internet
draft published, but ietf-types@iana.org is just
a review forum, not a place to send "please register this"
requests.

[... extensive discussion of IANA queues, IESG queues,
and RFC editor queues, and interactions between them ...]


Mealling and others recommended:

  1) Put the answers to the media type registration questions
	in the W3C REC-track spec (perhaps in an appendix).

  2) Write short internet draft (a "pointer document") saying
   "We want to register the mime type defined in this w3c
   spec appendix....". Make sure the "IANA Considerations"
	section clearly says
	"The IANA is directed to register the mime type 'foo'
	using the template found in the document bar
	upon publication of this document as an RFC."

    The answers to the ~7 MIME registry questions *don't* have to be
	in the pointer document.

  3) take the W3C REC-track document thru the REC track,
	updating the pointer document at least once
	every 6 months.

  4) Once the W3C REC-track document is done (i.e. is
	a W3C Recommendation), send a request to rfc-editor,
	asking that the pointer document be published
	as an informational RFC.

	The RFC editor will put it on the IESG queue.
	After IESG disposition, it'll be published
	as an RFC and, (as a result of the clear
	"IANA considerations" section from step 2)
	it'll go into the IANA registry.

Freed advised copying him at step 4, as he is often
in a position to expedite such requests.

Berners-Lee noted that updating the registry so late
in the process (a) provides no guarantee against
collisions, and (b) provides no lookup facility
via the IANA registry during development in 
the W3C REC track.

Mealling observed that the risk of collisions
(a) is mitigated by periodic publication of
Internet drafts, and (b) has yet to materialize
in any case.

Freed offered to follow up on the idea of
a provisional registry, where entries
would expire with the corresponding
Internet Draft.

Mealling described a web database for pending URN namespace
registrations which satisfied a lot of concerns among
those involved in registering URNs.

ACTION Mealling: send details of URN registration workflow tool.
	[in IRC, Mealling cited
	  http://uri.net/urn-nid-status.html
	adding
	"that is currently down but it should work in an hour
	ro two...."]


ACTION Freed: persue update to RFC2048.
  (with ideas such as std. facet, provisional registry)
  draft out for review (not Internet Draft) in
  yokohama timeframe.

ACTION Reagle: send his understanding of how to get a W3C language
registered as a media type. [to where?]


>   -- W3C WGs registering HTTP headers
>     Device Independece WG HTTP header (cf timbl)

POSTPONED due to lack of time.

Freed mentioned a possibly relevant draft by Nottingham.

>   -- URIs for MIME types
> 
>   cf
> 
>    Re: [uriMediaType-9] Dear IETF ...
>    From: Larry Masinter (LMM@acm.org)
>    Date: Thu, Jun 06 2002
>    http://lists.w3.org/Archives/Public/www-tag/2002Jun/0050.html
> 
>    TAG Finding: Mapping between URIs and Internet Media Types
>    http://www.w3.org/2001/tag/2002/01-uriMediaType-9

POSTPONED due to lack of time.


>  -- FYI: W3C TAG issues related to IETF specs
> 
>   The TAG hasn't decided anything about these yet, but fyi...
> 
>   HTTPSubstrate-16 : Should HTTP be used as a substrate protocol?
>     Does W3C agree with RFC 3205?
>   http://www.w3.org/2001/tag/ilist#HTTPSubstrate-16
> 
> 
>   RFC3023Charset-21 : Do all "shoulds" of RFC 3023 section 7.1 apply?
>   http://www.w3.org/2001/tag/ilist#RFC3023Charset-21

Connolly explained that the media types agendum above
follows a decision of the W3C TAG; noted the
above issues have been accepted for the TAG issues
list, but not discussed to the point of resolution.

[... discussion of interested parties getting notice when a TAG issue
goes from dormant to active/open ...]

Freed asked what W3C would expect if the W3C TAG decided
not to agree with a published IETF BCP. Perhaps it
could be obsoleted, suggested Berners-Lee. Freed noted
that would require support from the BCP's author, but
noted that a separate RFC could be published as
an answer to the first; Berners-Lee asked if conflicting
BCPs were prohibited, and Freed didn't think so;
in fact, publishing opposing views in RFCs has been
encouraged by the likes of Jon Postel.

Connolly solicited advice on how to get wider consensus
during W3C TAG deliberations; Mealling suggested
notifying the apps-discuss mailing list
  http://www.apps.ietf.org/mailing-lists.html
as discussion on issues begins.


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Friday, 30 August 2002 16:22:23 UTC