Web Services Addressing WG Teleconference -- 6 Nov 2006

Resolution: Minutes of last meeting approved

bob: section 5.2.1 of SOAP binding allows for other URIs as anon. did we mean that ?
... brief discussion and will then vote of this question

marc: might be better phrased as "do we still mean it"

dhull: do we have a specific textual change to vote on

bob: will rip out 5.2.1

tom: looked for source of the text - not clear that text was explicit in discussion

editors can't recall where text originated

<pauld> it's in our spec and went through CR-PR-REC

<pauld> so nit-picking on the scribing seems a little mute

paco: what will be the gain by doing anything

marcgoodner: last sentence of 5.2.1 and 5.2.2 is the text in question

guilty == remove text allowing other anon URIs

not-guilty == keep that text

pauld: errate, version 1.1, ?

<dhull> It's not like the text has much force to begin with. Wouldn't an erratum be enough?

bob: how we do it is not relevant, will do minimum necessary

<MrGoodner> I think errata is enough

<MrGoodner> it doesn't seem to impact the wsa namespace

formal vote:

IBM - not guilty

BT - guilty

Microsoft: guilty

Sun - guilty

Tibco - not guilty

Nortel - guilty

W3C - guilty

BEA - not guilty

CA - guilty

Fujitsu - guilty

Hitachi - guilty

8 guilty, 3 not guilty

Resolution:Text indicating that other specifications may define URIs with behavior similiar to wsa:anonymous shall be removed from the rec version of the SOAP binding

<plh> http://www.w3.org/2005/10/Process-20051014/tr.html#rec-modify

plh: the change might affect conformance

marc: agree with paco, change put other uri out-of-scope for WS-A spec but doesn't preclude it

<pauld> so the lack of a MUST means a change doesn't impact conformance, and could fall into "Corrections that do not affect conformance"

dhull: saw text as guidance, doesn't affect conformance

<plh> 2. Corrections that do not affect conformance

<plh> Editorial changes or clarifications that do not change the technical content of the specification.

<dhull> +1

bob: does WG feel change is not related to conformance

WG: assents

Resolution: Removal of language agreed in above resolution is determined to be an editorial change since it does not affect conformance

bob: part 2 deals with failure of WSDL binding to support usage promoted by text we have voted to remove
... close 33 without action ?

paco: what about the proposals

bob: do we all agree that we need to do something to compose better with policy ?

WG assents

bob: deal with cr33 on its basis and then look at proposals to better support policy

paco: don't want to close cr33 until we have dealt with policy proposals

mrg: only anon marker needs to be fixed

paco: do we have a specific proposal ?

http://www.w3.org/2002/ws/addr/cr-issues/Overview.html#cr33

bob review the issue text

dhull: anon=optional seems to leave the door open to using RM anon but really it just pertains to WS-A anon

<dhull> actually, it's anon=required that just pertains to WS-A anon. Optional leaves things open.

paco: markers as designed not useful, issue is not with text in soap binding

katy: have lost track of rx requirement

<dhull> A&P proposal speaks to a different issue

<dhull> RX requirement is actually already satisfied

<dhull> Doesn't think WSA anon definition is restrictive at all

katy: when soap binding came out it was possible for rx to use their anon uri but with wsdl binding we restrict that

gill: is it possible to define a policy that restricts to anon only and then expand that to allow rx anon ?

glen: yes

paco: such usage would contravene policy usage recommendations

glen: no it wouldn't

paco: yes it would

repeat

paco: going to have to ignore the marker because it is badly designed

marc: unclear which marker are badly designed - i think everyone agrees that the original formulation doesn;t work with policy but several options have been proposed - some by paco - are any of these ok in paco's opinion

paco: some are ok, disagreeing with glen proposal for rx to create an assertion that overrides the ws-a one

dhull: composability not limited to rx

<dhull> so that means we're going back to LC?

quick poll confirms that everyone agrees that current syntax is broken wrt policy

Resolution: markers are broken and need to be fixed for policy

mrg: we just need to fix it wrt ws-a anon

dhull: use case is indicating haw async response can be sent

paco: having same discussion as last week

bob: don't think ws-a has exclusive on back-channel, we need to find a way to get out of the way of future specs

dhull: need a way to say stuff like: "can use mail addresses in reply to"

<dhull> but it's useful to say "I can't do anythng *but* anon" (or "I can't do anything *but* email")

marc: going over ground we covered in Japan F2F. think we need a way to state "i can do ws-a anon" and "I can't do ws-a anon" but not have assertions that shut out others

mrg: not worried if our assertion shuts out others

dhull: think we have new information now that requires revisiting Japan decision

<dhull> +1 on relevance of policy

marc: use of policy brings level of expressiveness that a WSDL marker doesn't have, the two marker proposed are all we need to define, other specs can define others

dhull: in what way is our current markup policy unfriendly ? does policy have a "this and this but not that"

paco: second is an open issue with policy WG, first is that assertion qname should capture meaning not attribute value or content

<gpilz> +1

marc: volunteers to propose a solution

gill: volunteers to help

<bob> thanks for scribing, Marc

- DRAFT -

Web Services Addressing WG Teleconference

6 Nov 2006

Attendees

Contents

Summary of Action Items

Scribe.perl diagnostic output