Web Services Addressing Working Group Teleconference

15 Nov 2004


See also: IRC log


David Orchard
Paul Downey
Rich Salz
Tom Rutt
Robert Freund
Yin-Leng Husband
Greg Truty
Rebecca Bergersen
Steve Vinoski
Jonathan Marsh
Michael Eder
Mark Peel
Anish Karmarkar
Jeff Mischinsky
Steve Winkler
Glen Daniels
Marc Hadley
Arun Gupta
Hugo Haas
Martin Gudgin
Harris Reynolds
Mark Little
Davanum Srinivas
Jacques Durand
Eisaku Nishiyama
Francisco Curbera
Marc Goodner
Ales Novy
Jiri Tejkl
Philippe Le Hégaret
Mark Nottingham
Robert Freund


Chair reviewed the agenda

Minutes were posted for last week, Chair called for corrections.

RESOLUTION: Last minutes and the F2F minures from 11/8 were approved

AI review

Issues 17 & 16, 31, 08 are done

Upcoming f2f Meetings

Send message to Chair if remote participation in Redmond F2F is desired

Jan 17-19 will be in Melbourne

Logistics and registration will be out in the next few days

April F2F will be around April 15. Meeting after that would be in early June.

Working Drafts

Marc: ready to publish first working draft

Chair: Working drafts will be reviewed prior to publication

Discussions on the list this week, be concise and clear, goal is to vote for working draft next week.

Marc: There are minor editorial changes remaining to make

Chair: Editors, if changes are made this week, please publish summary of changes to the list

New Issues


Chair: Shall XML schema be used to describe the spec? Chair asked if there are any objections
... There were no objections.

RESOLUTION: XML Schema will be used to describe the specification where appropriate

Discussion wrt XML 1.1 or pseudo schema or text for normative description

Jonathan: proposal to drop xml 1.1 for use in the normative description, Rutt agrees

Rutt: suggests that schema and text could both be normative but would be difficult to sync.

<rsalz> it's actuall *3* things to keep in sync: prose, xsd, psuedo-schema

Hugo: proposed that the schema be normative for 1.0

<anish> could we just have prose and xml schema for 1.0?

<rsalz> yes, please. :)

<anish> iirc, xmlp decided against xml 1.1 as soap 1.2 became a rec before xml 1.1 and SOAP 1.2 already had schema as normative

mnot: straw poll to restrict use to xml 1.0

<Marsh> +1 for Microsoft

+1 for Hitachi

<rsalz> +1 for datapower

<anish> i'm sypathetic to the complexity issue (look at wsdl), but perhaps preventing xml 1.1 may be going to far

<Greg> +1 for IBM

<RebeccaB> +1 for IONA

<mpeel> +1 for Novell

<TomRutt> +1 Fujitsu

<marc> +1 for Sun

<stevewinklr> +1 SAP

<TomRutt> How about restricting to the soap binding?

Straw poll was taken, and decision was made to restrict use to XML 1.0

RESOLUTION: WS-Addressing's use of XML will be restricted to XML 1.0

<scribe> ACTION: Rich to begin work on schema

<pauld> existing ws-addressing schema: http://schemas.xmlsoap.org/ws/2004/08/addressing


Assigned to Rebecca, Rebecca reviews verbally

Rebecca proposes adding one property to the epr where the complete wsdl may be located

<vinoski> +q

Chair: discussion to continue on this topic via the list

<Zakim> Marsh, you wanted to ask about interaction with [service-port] and/or [selected port type]

<TomRutt> +1 steve

anish: concern that if we go down this path, it may have implication with other issues.

Chair: continue discussion on the list



hugo: discussion is around uri cum reference properties exacerbating problems that exist with uris as they stand
... simplicity argument is that reference properties are known to the provider and are not part of the externally usable (to the service provider) address
... disadvantage is that to secure a uri, then it must be secured in its entirety

hugo: Web arch says that resources are uniquely identified by uris.
... not ready to make a conclusion

Chair: wants to kep it moving along

<stevewinklr> +1 to clear terminology

Rutt: in order to reconcile spec with web architecture, then it will be necessary to refine the name to be something that will help foster understanding

<anish> +1 to marc

Marc: analagy has been made that reference parameters are a lot like cookies, and that cookies are not used for addressing

Chair: from the web architecture, cookies are broken, and that they are used to carry addressing information (as well as other info)

issue 3

Chair: gudge not online

<rsalz> away away from my desk for 30 minutes, in case you care

<anish> +1 to hugo, this is tied with the issue currently being discussed in WSDL as to -- what is a node?

hugo: node is a logical entity that may be addressed uniquely

Chair: possibility is to accept the proposal and say that this is the list of starting point MEPS that will be covered
... asked for objection, none heard i003 is closed with gudge's proposal

and that MEPS covered will be part of our continuing discussions.

RESOLUTION: i003 is closed by accepting gudge's proposal for list of MEPs; supplied text will be starting point for further work.


Marc: Several questions arised related to the scope of reply-to
... and with which MEPs reply-to had what meaning.
... bigger issues need to be resolved prior to i028's resolition


Relates to semantics of Fault-to

Marc: ? must we define the implications of the presence or absence of each of these for each MEP?

Chair: may be part of a larger issue which is conformance to each of the documents of the specifications
... will try to raise this issue this week.


Chair: propose closure to i012 with proposal rev 2 by paco.

<Chair> http://www.w3.org/mid/OF56AF0FA7.74E5CA56-ON85256F46.00761383-85256F47.0042B98B@us.ibm.com

Chair: hearing no objection, i012 is closed with proposal 2

RESOLUTION: i012 is closed with proposal 2


hugo: wsdl 1.1 and wsdl 2.0 descriptions may both exist. What would then be the interprety of the action property

Chair: going forward, we may consider splitting hugo's sub-issue of default actions and closing the rest

marc: if action were optional, this issue would go away

anish: this is a bigger issue than action since we have reference to wsdl in other areas

Chair: let us split this up and record this as a new issue, can we close the remainder of i019?
... asked for objections/ hearing none, i019 is closed with the carve-out of the new issue to be posted

RESOLUTION: i019 is closed by accepting Hugo's proposed changes, except for those relating to the default action.


hugo: we have a choice of making addressing ubiqutous, and we could request that address be defined in the soap bindings.
... addressing ought also to be represented in tools. In order to do this, one ought to define a soap 1.2 module

<anish> +1 to option #2

as well as a proposal to cover in WSDL 2.0

<marc> -infinity to option #1

<pauld> seems like ws-addressing is a good pilot for WSDL 2.0 extensibility .. so option#2

<GlenD> +1 to option 2

+1 to option 2

<GlenD> (it'll take an awful lot of +1's for option 1 to make up Marc's vote....:))

<anish> :-)

just calcualte the average vote in favor

Chair: WG will go forward with option 2.

Marsh: Believes issue needs more work concerning use over all versiosn of wsdl
... wsdl 1.2 offers one extension mechanism, 2.0 offers two. Believes debate would be meritotious as to which mechanism should be used.

Hugo: it was his intent to cover all of our bases in wdsl

Marsh: we could develop a proposal for the use of both extension models so that we could see the proposals side by side.

<scribe> ACTION: hugo: Will send another email with a refined proposal for i021 by Nov 26


Multiple ports in EPRs

Chair: (aside) new issue, pls get involved in new multiple port and EPR issue.

issue number to be assigned

<scribe> ACTION: anish: Nov 17 to explore the scope of the action defaults issue

return to i026

steve: Believe that a lot of services are accessible over multiple transports/protocols. What we are finding is that customers would like to re-use their legacy protocols.

trying to put together a true service reference with one construct. Alternative is presented to include optional information to the EPR so that all of its addressing capabilities could be specified

Additional alternative was described that the scribe missed completely

pls refer to email proposal for detail

steve: proposed that complete addressing information be included so that network operation to look up further addressing information would be unnecessary

<Zakim> Marsh, you wanted to ask where the boundary between these proposals and the whole WSDL is.

Chair: Is this isue complementary with i033?

steve: some applications have a lot of pre-configured knowledge of addressing situation, other applications have little pre-configured knowledge.

rutt: would reference properties be the same for these different situations?

<anish> if we go down this path -- then i would prefer option 1 over option 2

Chair: issue needs more seasoning.

Rutt: understanding of i001 will help us figure this out.

Rebecca: askes for activity re i023 on the list

<scribe> ACTION: anish send a mail to rebecca to start by 11/18

Summary of Action Items

[NEW] ACTION: anish send a mail to rebecca to start by 11/18 [recorded in http://www.w3.org/2004/11/15-ws-addr-irc#T22-01-07]
[NEW] ACTION: anish: Nov 17 to explore the scope of the action defaults issue [recorded in http://www.w3.org/2004/11/15-ws-addr-irc#T21-43-12]
[NEW] ACTION: hugo: Will send another email with a refined proposal by Nov 26 [recorded in http://www.w3.org/2004/11/15-ws-addr-irc#T21-40-52]
[NEW] ACTION: Rich to begin work on schema [recorded in http://www.w3.org/2004/11/15-ws-addr-irc#T20-38-52]

Minutes formatted by David Booth's scribe.perl 1.95 (CVS log)
$Date: 2004/11/20 00:49:06 $