Minutes 20050110

Those minutes have not been reviewed by Mark. If you're not
appropriately listed in the attendance, let me know directly and I'll
fix the online version.

Available at:
  http://www.w3.org/2002/ws/addr/5/01/10-ws-addr-minutes.html


Text version:

   [1]W3C

      [1] http://www.w3.org/

                   Web Services Addressing WG Teleconference

10 Jan 2005

   [2]Agenda

      [2] http://www.w3.org/mid/5EA8B6AE-6114-11D9-9594-000A95BD86C0@bea.com

   See also: [3]IRC log

      [3] http://www.w3.org/2005/01/10-ws-addr-irc

Attendees

   Present
          Yin-Leng, Steve, MarkL, MarkN, MarkP, David, Arun, Philippe,
          Bob, Jonathan, Marc, Tom, Greg, Glen, Ugo_Corda, Anish, Paco,
          Hugo, Rebecca, Pete, Paul, Harris, Umit, Michael, Rich,
          Davanum, Jeff

   Chair
          Mark Nottingham

   Scribe
          Mark Peel

Contents

     * [4]Topics
         1. [5]Review of action items
     * [6]Summary of Action Items

     _________________________________________________________________

   Agenda: Minutes, action items (1 by 1 due to upcoming F2F)

   Next week's meeting, 1 last issue, issues list; no other business

   No corrections to minutes; minutes approved.

   Jan 3 minutes.

Review of action items

   Issue 21: Anish will take a look and confer with Hugo before F2F; due
   Friday

   Mark N finished his action item about WSDL

   Paco finished reviewing core

   Scope of addressing properties incorporated in drafts

   Gudge to redraw charts; not present.

   Hugo reviewed interface names in submission; drop this item.

   Jonathan taking over MarcH's action to make proposal re: issue 35; due
   Friday.

   Thanks for the correction.

   Future meetings: F2F next week; any problems/suggestions for agenda?
   None.

   No questions on logistics.

   Issue 26, multiple ports, deferred because owners not attending F2F

   Phone access by Zakim

   Joint meeting with WSDL working group: 2 hours on Wed.

   What are the issues; where do these specs affect each other?

   <anish> another two issues that are related to what WSD WG is doing
   are: 20 and 21

   Issue 22; bindings of WS-Addressing and SOAP to WSDL unclear

   Jonathan to list options to frame discussion; Mark N: may not be time
   to address all of this

   ACTION: paul to write up the problems this WG has had with WSDL
   extensibility

   What was resolution of use of WSDL 1.1 location feature? The belief
   was that there was no objection, but that will be cross-checked

   MarkN: Should someone prepare of list of issues with WSDL already
   discussed?

   ACTION: davido to put together a presentation about the relationship
   of addressing to wsdl wrt to MEP/binding/variable-adressing

   David's proposal will be discussed at F2F; may discuss with WSDL group
   afterward.

   Perhaps Issue 20 should also be discussed with WSDL group

   Run-through of changes to editor's drafts since last F2F; scheduled
   for half an hour

   WSDL joint meeting can't be moved to Wed morning for benefit of
   phone-ins

   Issues: First Issue 9; Paul completed his action item.

   Paul's walkthrough: can we allow multiple ReplyTos using mustIgnore?

   There's no ordering on the ReplyTos, hard to tell which to ignore

   Add annotation against ReplyTo to collate? Might cause more issues

   Paul prefers option 3; he prefers a processing style, but wishes to
   consult group

   Tom Rutt: prefers option 3

   Jonathan: How does processing style attribute work?

   Paul: this is one of the problems

   Jonathan: Is there a difference between a header I don't understand
   and an attribute I don't understand?

   Bob wants multiple ReplyTos to as alternative endpoints

   TomR: this spec shouldn't describe the semantic of more than one
   ReplyTo

   Paul: still prefers Option 3 after discussion, but not convinced of
   processing style

   <TomRutt> I like option 3, without processing style, with the "if more
   than one, this spec does not define behaviour"

   Paul: 1 processing style for whole core; different replyTos would have
   the same processing style

   <anish> i kinda like option 2

   Umit: can we discuss the other options?

   MarkN: let's, but wanted to discuss 3 first

   <anish> but would like to understand the full semantics of
   'processingStyle'

   MarkN: do WSDL bindings constrain cardinality? Paul: no, wide open;
   WSDL constrains in Option 2

   Paul: processing style is really a separate proposal

   ACTION: paul to put together a separate processingStyle proposal
   (issue 009)

   Hugo: Paco's amendment to his proposal just fine.

   Issue 14 closed? No objections.

   Issue 34, action defaults: Jonathan proposed porting WSDL 1.1 to 2.0

   Hugo had a proposal; should we wait on Issue 35 before closing 34?

   Hugo: it is a shame we are not going re-use a WSDL 2.0 spec URI;
   instead we are introducing a new mechanism.

   We should communicate this to the WSDL group.

   Jonathan proposes a syntax taking target namespace and appends
   interface name and a slash etc.

   It's backward compatible with WSDL 1.1

   Close Issue 34 with Jonathan's proposal? No objections

   Issue 35: Jonathan will get this to us by end of week

   Harris: can send out template for test cases to list today.

   Issue 25: multiple actions. Postponed due to Gudge's absence.

   Issue 40, processing model for SOAP faults:

   Arun has sent a message to list to start discussion.

   Issue 38: Gudge owns, Paco has completed an action item.

   Paco's proposal changes core spec; marc's changes are in WSDL binding;
   changes are complimentary? yes

   Marc: my proposal is a relatively complete resolution

   for Issue 38

   People still need to look over Issue 38 before closing.

   Marc has added editor's notes to aid review.

   Issue 6, property optionality:

   Issue 6, part 1, will be deferred to F2F to give MS further time to
   review

   Issue 7, processing model of headers in the SOAP binding

   Touches on 21 and 22; and also optionality issue

   Marc: we define SOAP headers but not what a receiver should do with
   them

   Messages traverse multiple hops; intermediaries should process
   WS-Addressing headers

   There's an assumption in the spec headers are targetted to ultimate
   recipient, not intermediaries

   Should we specify re-inserting headers when intermediaries process?

   Should headers be targetted to Next?

   Marc is not sure they should be mustUnderstand=true

   Paul: what are use cases?

   Anish: would it make sense to divide properties into buckets, some
   which must survive to ultimate recipient, some which don't?

   Marc: perhaps, but must consider; he believes most must survive

   MarkN: doesn't the spec say all must be end to end? Marc agrees

   <MSEder> queue

   Marc: use case is intermediary doing routing based on WS-Addressing
   headers

   GlenD: 2 intermediaries, known to sender, transparent to sender

   <anish> I assumed that Marc was talking about forwarding
   intermediaries

   <anish> as wrt active intermediaries all bets are off

   does your thinking include both types of intermediaries?

   Marc: we could decide every intermediary is (or may be) an active
   intermediaries

   MarkN: Marc, take this to email. Marc: okay

Summary of Action Items

   [NEW] ACTION: davido to put together a presentation about the
   ... relationship of addressing to wsdl wrt to
   ... MEP/binding/variable-adressing
   [NEW] ACTION: paul to put together a separate processingStyle
   ... proposal (issue 009)
   [NEW] ACTION: paul to write up the problems this WG has had with
   ... WSDL extensibility

     _________________________________________________________________


    Minutes formatted by David Booth's [7]scribe.perl 1.95 ([8]CVS log)
    $Date: 2005/01/11 03:22:58 $

      [7] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribe.perl
      [8] http://dev.w3.org/cvsweb/2002/scribe/scribe.perl