W3C XML Protocol teleconference, 11 July 2001
1. Roll call
Present 34/30
- Canon Jean-Jacques Moreau
- Data Research Associates Mark Needleman
- Active Data Exchange Richard Martin
- Akamai Technologies Mark Nottingham
- Compaq Yin-Leng Husband
- DaimlerChrysler R. & Tech Mario Jeckle
- Engenia Software Eric Jenkins
- Epicentric Alan Kropp
- Ericsson Research Canada Nilo Mitra
- Fujitsu Limited Kazunori Iwasa
- Hewlett Packard Stuart Williams
- IBM Doug Davis
- IBM David Fallside
- IBM John Ibbotson
- IDOOX Jacek Kopecky
- Intel Randy Hall (scribe)
- Jamcracker David Orchard
- Macromedia Glen Daniels
- Matsushita Electric Ryuji Inoue
- Microsoft Corporation Henrik Nielsen
- Mitre Marwan Sabbouh
- Mitre Paul Denning
- Novell Scott Isaacson
- Philips Research Amr Yassin
- Rogue Wave Murali Janakiraman
- SAP AG Gerd Hoelzing
- SAP AG Volker Wiechers
- Software AG Michael Champion
- Sun Microsystems Chris Ferris
- Tibco Frank DeRose
- Unisys Nick Smilonich
- Vitria Technology Inc. Tony Lee
- W3C Hugo Haas
- W3C Yves Lafon
- Xerox Ugo Corda
Excused
- Active Data Exchange Shane Sesta
- Canon Herve Ruellan
- Compaq Kevin Perkins
- DaimlerChrysler R. & Tech Andreas Riegg
- Engenia Software Jeffrey Kay
- Epicentric Scott Golubock
- Fujitsu Limited Masahiko Narita
- IDOOX Miroslav Simek
- Intel Highland Mary Mountain
- Macromedia Simeon Simeonov
- Microsoft Corporation Paul Cotton
- Philips Research Yasser alSafadi
- Rogue Wave Patrick Thompson
- Software AG Dietmar Gaertner
- Sun Microsystems Marc Hadley
- Unisys Lynne Thompson
- Vitria Technology Inc. Richard Koo
Regrets
- Library of Congress Ray Denenberg
- AT&T Mark Jones
- Commerce One Jay Kasi
- DataChannel Brian Eisenberg
- DevelopMentor Martin Gudgin
- Informix Software Charles Campbell
- Interwoven Mark Hale
- IONA Technologies Oisin Hurley
- Lotus Development Noah Mendelsohn
- OMG Henry Lowe
- Oracle David Clay
- Propel Daniela Florescu
- WebMethods Randy Waldrop
Absent
- AT&T Michah Lerner
- BEA Systems Dan Frantz
- BEA Systems Jags Ramnaryan
- Bowstreet Alex Ceponkus
- Bowstreet James Tauber
- Commerce One David Burdett
- DevelopMentor Don Box
- Group 8760 Dick Brooks
- Informix Software Soumitro Tagore
- Interwoven Ron Daniel
- IONA Technologies Eric Newcomer
- Library of Congress Rich Greenfield
- Netscape Ray Whitmer
- Netscape Vidur Apparao
2. Agenda review, call for AOB
David Fallside: Reviewed the agenda which follows.
1. Roll call, scribes for minutes/action items (1.00 + 5m)
RandyH (Intel) is scribe for minutes, Action Items taken on
IRC (Henrik et al)
2. Agenda review, call for AOB (1.05 + 5m)
4. Review action items, (1.15 + 10m)
5. Pseudo-Review of XML Encryption (1.25 + 5m)
6. XMLP WG schedule (1.30 + 25m)
-- Work items:
Infoset, Primer
(Transport Bindings), RPC, Encodings
Resolve outstanding issues
Check requirements met
Validate usage scenarios
Validate conformance criteria
-- Meetings:
Can the WG agree to (1) a 3-day f2f on November 27-29 on the east
coast USA (modulo acceptance by XML Schema, XML Query and XSLT), and
(2) a 3-day f2f in mid-February in France to co-incide with the XML
Plenary?
7. Transport Binding Architecture (1.55 + 30m)
Status report from Protocol Binding TF members on (copied from last
minutes
8. Any Other Business (2.25 + 5m)
There was 1 AOB Item addressed during agenda review: Scott Isaacson
confirmed dates for September F2F: 11-Sep-2001 through 13-Sep-2001 (3
days).
4. Review action items
2001/03/28: David Clay to Start working on reconsiliation between
what he has proposed with the most up to date version of the AM and
Henrik's glossary mapping.
Discussion of this pending action item was postponed since
David Clay was not present.
2001/06/06: Gudge and Henrik - Get back to the author of issue 19 -
bring the issue on xml-dist-app.
This item has not been taken to xml-dist-app and is still
pending.
DavidF - This action item (2001/06/06) should be morphed into
revisiting the issue on xml-dist-app.
2001/06/07: Hugo - Start thinking about how to record conformance.
Hugo: this item is pending.
2001/06/20: PaulC -PaulC to send e-mail on issue 30 by July 11
DavidF: this item is pending
2001/06/20: DavidF - Think about whether we need the block/header
entry distinction issue on the list
DavidF: thought he had put this on the issue list. He will
check, and so the item remains pending
2001/06/27: DavidF - Write an email about the HTTP-Ext section with
ChrisF and refer to issue # in issues list.
Chris has just sent the draft e-mail to DavidF who has not yet
had time to address (pending)
5. Pseudo Review of XML Encryption
HugoH: Hugo sent e-mail to list looking for volunteers. David
Orchard did a review. Hugo will incorporate David Orchard's feedback
and work with Joseph Reagle. Hugo invited more feedback and noted
friday deadline
DavidO: If people disagree with feedback, it may be appropriate to
come up with a formal working group response
HugoH: Joseph didn't want a formal reply from WG but more from
individuals. If others have input, please send them to Joseph. Hugo
suggested that David O. discuss with Joseph at the Processing Model
Workshop (12, 13-July-2001).
6. XMLP WG Schedule
DavidF: David sent a detailed timeline to WG members on Saturday, and
proposed we briefly go through this propsal then address questions.
Group agreed it would be useful to go through the e-mail.
DavidF presented the schedule as follows:
The schedule starts with Working Draft publications that happened on
9-July. The next item is September F2F in San Jose. David noted that
this working group seems to be able to publish after having a F2F. So
the next milestone for publishing would be end of September (after the
f2f).
David proposed we have another F2F meeting in late November. There is
a possible slot back-to-back with XSL, Schema, and Query WG's. We
could have F2F in this timeframe. Following the same sequence of
events as September, we would publish again after this F2F. This
working draft will be the Last Call WD. Last Call is a significant
event in lifespan. When a working group publishes a Last Call, it is
indicating that the members believe they have met requirements in both
the charter and the requirements document and that the membership
believes all open issues have been addressed. It's like saying "We
think we're done." This is a significant milestone.
From late November-early December through February, we will be in Last
Call review which gives the WG 3 months to solicit comments and
respond to them. We would give people approximately 1.5 months (out
of the 3) to send in comments. In responding to comments, we must be
particular in how we respond: there needs to be a paper/URL trail.
When we believe we have responded to comments from Last Call review,
we put together another version of the WD (Proposed Recommendation)
that goes to the Director and all members of the consortium for final
review. Again, there may be issues raised against the PR. Timeframe
for Proposed Recomendation feedback is late Feburary to the end of
April.
This would mean that we would go immediately from Last Call to PR
without a Candidate Rec stage. CR is used for specs with little
implementation history. The point of CR is that at the end, you need
to be able to show implementations. This will apply to us, but given
that there are already a large number of SOAP implementations out
there today, we will simply have to show the Director that the
technical features of the PR have been implemented.
GlenD: Noted that several changes from the SOAP 1.2 working draft
have already been implemented in SOAP::Lite.
DavidF: It would be helpful if Glen could send a pointer to
xml-dist-app that someone has done this. We need to have this audit
trail.
GlenD: Agreed.
DavidF: In August and October, David has listed possible WDs. These
are suggestions to the group that as we do significant pieces of work,
it is better to get them out for feedback sooner rather than later.
David thinks we should target one in September and at least one more.
Does this seem reasonable?
Mark Nottingham: Is the November F2F something that works for
members.
DavidF: We haven't discussed but it's on the agenda later. First we
need to determine whether this is a good approach.
Silence is assent so the proposal and schedule as described above is
accepted.
DavidF: Let's move to F2F discussion. Agenda ordering changed,
Meetings
The chair proposed a 3-day F2F November 27-29 on the east coast.
XSLT, Schema, and Query are scheduled to meet Monday through Friday
after Thanksgiving. If they stay with that arrangement, and given the
overlap in the groups, it would make sense for us to meet during the
following week, (December 5, 2001) or we could meet week of
26-November 2001.
The chair proposed making the meeting Tuesday, Wednesday and Thursday
to accommodate travel. XML Query and XML Schema has indicated that
they would like to change the date of the meeting so that it is the
week of December 5. If this is the case, would people be able to make
a meeting on 26-Nov through 28-Nov on East Coast.
MarkN: both weeks are difficult
GlenD: If this were in Boston, either week is OK; otherwise could not
attend a meeting during the week of 26-Nov.
DavidF: We do not have venue yet.
Group agreed that we could meet either the week of 26-November or the
week of 3-December depending on when XSL, Schema and Query are
meeting.
DavidF: Will negotiate with Schema, Query and XSLT for the date.
The chair also proposed a F2F in France in mid-February to coincide
with XML Plenary.
There was no objection to this proposal so we agree to have another
F2F coinciding with XML Plenary in France.
Work Item: Primer Discussion
The chair believes a primer is something we need to have.
Mark (Hale?): Schema primer was a lot of work.
Nilo: Suggested 2 ideas. The first is to do something along the
lines of what Tim Bray with Annotated XSL: take the spec and explain
motivation for specification feature. The second suggestion to
develop a users guide.
Mark Nottingham: Doesn't know that the WG should pursue this due to
resource constraints. This would certainly add to our work.
Henrik: Agrees with Mark Nottingham. It would be nice to have
descriptions of how we build on top of the specifications.
Mike Champion: Would have had a hard time getting adoption without a
primer.
Henrik: Most agree that XML Schema is very difficult. SOAP is not in
the same category.
Nilo: Believes we need to validate the specifications against use
cases. Doesn't know how we can validate without putting these in a
primer. This would be a guide to how application developers could use
SOAP.
DavidF: We might use some of our work on scenario validation in the
primer.
DavidF: As the author of Schema primer, he indicated that it was not
a bad use of his time. In working through the primer, he encountered
a couple of situations where he uncovered issues that the spec authors
didn't see. The objection thus far seems to be resource commitment.
David noted that if someone handed a primer to the WG, we wouldn't say
no.
Nilo: Nilo has come across a lot of tutorials but people are not able
to see range of usages without tutorials. Nilo is willing to draft a
proposed primer and distribute to the WG.
JohnI: John is willing to help with primer as well. There will
always be a number of people to pick this up from scratch.
DavidF: Can we agree to sanction Nilo with help from others to
produce a draft primer.
Henrik: Asked a question: what is scope of this? Is this a
discussion of how to write modules? If this is a discussion of how
SOAP fits into model mechanisms, it will take a lot of time from the
WG because the WG will have to show that the scenarios do, in fact,
fit the model.
Nilo: The point of primer would be to outline abstract items that
aren't described in detail spec.
DavidF: Nilo can you write a proposal for primer and what it
contains?
Nilo: Yes
WG agrees to ask Nilo and others to generate a proposal outlining,
with some examples, the Primer.
Work Item: Infoset Description Discussion
DavidF: Martin Gudgin provided a rewrite of section 4 in terms of
Infoset. We need to decide whether we're going to expand on this.
Mark Nottingham: There are 2 issues--whether we use infoset to
describe structures and whether we use Infoset to describe
serializations in the future. Maybe we should discuss technical
issues first.
GlenD: I don't think we're considering doing the latter yet.
[AlexC left telcon]
DavidF: Noah had suggested that a normalized XML serialization that
meets infoset is something that may be addressed by a protocol
binding.
GlenD: Noted that this may or may not use angle brackets.
DavidF: Is there anyone who thinks the use of infoset to describe
structures is not a good idea.
There was no objection
DavidF: We will provide an XML Schema that defines a serialisation,
but we won't provide other serializations. To Mark Nottingham - does
this obviate your concerns?
Mark Nottingham: Yes
DavidF: Proposed that we request that Martin Gudgin continue
developing infoset description of SOAP
Henrik: To be specific - for the envelope.
DavidF: I believe this is true.
Mark Nottingham: Henrik is alluding to the possible use of Infoset in
a protocol binding.
Henrik: Yes. I think we should say that the Infoset describes the
envelope, and the protocol binding is a different discussion.
ChrisF: Suggested that Martin might attempt to provide the same
characterization of the http binding so we can get a sense of what it
would look like.
DavidF: We are postponing the discussion of how to deal with the
protocol binding.
DavidF: We have decided to provide an infoset description to describe
the envelope.
DavidF: Chair takes Action Item to ask Martin to complete basic
description using Infoset.
DavidF: Noted that have finished with first 2 work items under XMLP
WG schedule
Transport Bindings, RPC, Encodings
DavidF: Noted that we have kicked off a transport bindings Task
Force and proposes creating RPC and Encodings Task Forces. David is
not suggesting we kick off RPC and Encoding Task Forces immediately
because it may dilute transport bindings work. David will produce
more detailed schedule of how these task forces play out over time.
David noted that serializing will not allow us to complete on time.
Chair proposes staggering start time.
No objection
Frank DeRose: Volunteers for the RPC task force.
DavidF: Requests that working group members volunteer when we make
the call for participation. Does it make sense to stagger start times
of task forces?
No objections were raised.
Resolving Outstanding Issues
DavidF: Another work item for the WG is to resolve all outstanding
issues. He indicated that he will be soon be bringing a list of issues
to the WG that are believed to be closed, so that the WG can verify
this is indeed the case. We will also need to validate that we meet
usage scenarios and requirements.
DavidF: On section 2 of processing model, we did a lot of
clarification work. He believes we will do the same with RPC and
encoding sections. This should take place within the context of the
task forces.
7. Transport Binding Architecture
DavidF: Mark Nottingham and Mark Jones posted descriptions, and
StuartW posted a detailed description of a scheme for transport
bindings. The TF needs to work on bringing these together. The chair
requested a brief discussion of the 3 proposals followed by discussion
of recommendations from the working group for the task force to come
back with a single proposal.
DavidF: Asked Stuart to describe his proposal.
Stuart: Proposed putting something together something based on work
by John I and Marc Hadley. In his proposal, Stuart attempted to
define a purpose and architecture for binding. The discussion of
materials didn't get much beyond framing a process for a binding. He
indicated that this should not be considered a polished proposal.
Henrik: I agree that the main argument is the definition of the
purpose for the protocol binding.
Stuart: The discussion hasn't proceeded beyond defining the purpose
for the protocol binding.
GlenD: there does not seem to be consensus.
Stuart: There were 2 forms of wording. Purpose of the binding is
there to describe how to make use of an underlying protoocl to
transport a message.
Henrik: The next question is what do we need. Should it discuss
message exchange patterns, should it describe QoS? Henrik does not
believe we should go beyond how we stick a SOAP message inside
something else.
Chris: It's not just sticking something into something else. We
cannot just say, "you stick something in HTTP" without specifying how
you return response.
Henrik: There needs to be a description of message exchange and
semantics. It doesn't appear to be a property of binding to say
whether you expect to see response for a message.
GlenD: What you are suggesting is an RPC abstraction, HTTP
abstraction and semantic bindings
DavidF: Henrik - what do you think the SOAP 1.1 HTTP binding actually
does describe?
Henrik: We have http messages that can contain SOAP message. This is
how we stick a SOAP message into a request (and response). The HTTP
does not require that you get a SOAP response back. You may get a 202
error. At the end of the day, HTTP is a request/response messaging
pattern. If we show how to break HTTP, we will get a lot of flak for
that.
DavidF: We are out of time. The protocol binding TF is charged with
solving the problems.
GlenD: Asked for a Transport Binding telcon to address the situation.
DavidF: Would like to have someone lead this group. Please send
e-mail to David if you wish to lead. Need to convene a telcon in a
short period of time. Chair asked if there were objections to a phone
call happening sooner than the W3C process normally allows.
No objections were raised.
Discussion of setting up Protocol Binding telcon.
DavidF: We will have to have one Friday (late afternoon) and possibly
one Monday. Request to Henrik to set up telcon for 4pm Eastern time.
8. Any Other Business
No additional items.
Meeting adjourned at 1:35pm PT