See also: IRC log
<bob> Scribe: Gil
Bob: agenda bashing
... F2F planning around Cannes
<anish> Agenda: http://lists.w3.org/Archives/Public/public-ws-addressing/2006Feb/0021.html
<bob> http://www.w3.org/2002/ws/addr/6/01/30-ws-addr-minutes.html>
Johnathan: would like to keep "2005-11-28: i059" open
Mark Hadley raised a concern with side effects of resoulution to cr17
Mark is not on the call
Anish: wonders if David Orchard will be presenting a re-worded issue?
Bob: I haven't seen the revised proposal. I will contact David about that.
<agupta> I see it correctly though
All: Agreed that we should defer this discussion until Marc is on the call.
Katy, Umit, and Jonathan speak in favor of taking WSDL Binding doc to last call
<anish> +1 to getting to LC quickly
Tom doesn't mind going to LC if this issue doesn't effect markup
RESOLUTION: take the WSDL Binding document to LC
<anish> http://www.w3.org/2002/ws/addr/cr-issues/#cr18
<anish> we have 2 proposed resolutions: one from DavidH and one from Paco
<bob> d hull's proposal: http://lists.w3.org/Archives/Public/public-ws-addressing/2006Feb/0024.html
<bob> Paco's proposals: http://lists.w3.org/Archives/Public/public-ws-addressing/2006Feb/0025.html
<anish> so, immediatedestination talks about the next hop, whereas wsa:To talks about the 'final destination'
<anish> if this is correct, then saying that anon means immediatedestination would not be right.
<anish> david, I *think* i understand what you mean (about anon and multi-hop)
<anish> since anon is defined by the binding, and bindings address MEPs which define what immediatedestination is, we are ok
<anish> is that roughly right
<Zakim> dhull, you wanted to say that anonymous won't work for multi-hop
<pauld> i made the point that option#3 clarifies the status quo, adding semantics to an anonymous To will impact at least one implementation good enough to participate in CR testing and would therefore be a "substantive change"
<Zakim> anish, you wanted to see if there was any support to make [destination] optional
<GlenD> +1 to making it optional
<dhull> FWIW: Option 4 is at http://lists.w3.org/Archives/Public/public-ws-addressing/2006Feb/0026.html, option 4' is at http://lists.w3.org/Archives/Public/public-ws-addressing/2006Feb/0028.html
<Jonathan> -1
<GlenD> Destination is tricky, because that gets into dispatch ideas. Don't make assumptions.
<GlenD> i.e some bindings/implementations don't need a WSAddressing [destination] property, just done by the transport
<uyalcina> my concern is if it is not mandatory, we would have to define when we need it. I have a problem with that.
<uyalcina> I hate contextual semantics.
<GlenD> Why would you need to define when you need it? You have an EPR, either from a WSDL or from another message. If that EPR tells you to put in a <to>, you do it. If not, don't.
<GlenD> If the <soap:address> is sufficient, then you don't need <wsa:To>, right?
<vikas> anonymous destination could mean "Hey DNS find me the nearest Gateway"
<anish> another discomfort I have is that we define the default in the core (which is anon) and don't say what it means
<GlenD> ah, I guess there's always an <Address> in an EPR, eh?
<dhull> +1 to discomfort over undefined defaults
<bob> 4' http://lists.w3.org/Archives/Public/public-ws-addressing/2006Feb/0028.html
<dhull> http://lists.w3.org/Archives/Public/public-ws-addressing/2006Feb/0026.html
<bob> opt 3. Middle of the road approach: retain the defaulting of the To header to
<bob> > anonymous, but re-state that its use is actually dependent on the
<bob> > interpretation that the transport binding gives to the anonymous URI. Add a
<bob> > note indicating that for the SOAP/HTTP case the anonymous URI is only used
<bob> > to indicate the use of the HTTP back channel so it can only be used in
<bob> > reply messages.
<pauld> chad, question approaches for finality of this issue
<pauld> chad, question: options for CR18
<pauld> chad, question: options for CR18 and CR20
<pauld> chad, option 1: option 1
<pauld> chad, option 2: option 2
<pauld> chad, option 3: option 3
<Jonathan> chad, option 3: Paco's option 3:
<anish> option 4': make [destination] optional and remove defaulting to anon
<Jonathan> chad, option 3': Paco's option 3 but limited to SOAP 1.1
<TonyR> chad, list options
<dhull> option 4: use ImmediateDestination where defined (request at the least), InboundMessage for response. Special case SOAP1.1/HTTP as per Paco's 3.
<bob> opt 3: Middle of the road approach: retain the defaulting of the To header to
<bob> > anonymous, but re-state that its use is actually dependent on the
<bob> > interpretation that the transport binding gives to the anonymous URI. Add a
<bob> > note indicating that for the SOAP/HTTP case the anonymous URI is only used
<bob> > to indicate the use of the HTTP back channel so it can only be used in
<bob> > reply messages.
<Jonathan> chad, option 3: Paco's option 3:
<Jonathan> chad option 3a: Paco's option 3 but for SOAP 1.1 only.
<Jonathan> chad, option 3a: Paco's option 3 but for SOAP 1.1 only.
<pauld> chad, options?
<Jonathan> chad, drop option 1
<Jonathan> chad, drop option 2
<dhull> chad, options?
<Jonathan> chad, option 4a: make [destionation] optional and remove defaulting to anon
<Jonathan> chad, option 4: DaveH's
<Jonathan> chad, option 4: Use SOAP 1.2 ImmediateDestination property
<Jonathan> chad, options?
<anish> chad, list options
<dhull> "use SOAP f&p"
<dhull> chad, options?
<pauld> vote: 3
<GlenD> vote: 4a, 4, 3a
<dhull> vote: 4
<uyalcina> vote: 3
<Jonathan> vote: 3
<David_Illsley> vote: 3
<Paco> vote: 3
<vikas> vote: 4
<anish> vote: 4a
<TRutt> vote: 3, 4a
vote: 4a, 4, 3a
<Katy> vote: 3
<bob> vote 3
<TonyR> vote: 4a, 3
<dhull> vote: 4, 3a
<agupta> vote: 4a
<Nilo> 4a, 3a, 3
<hugo> vote: 4, 4a, 3, 3a
<PaulKnight> 3, 4a
<PaulKnight> vote: 3. 4a
<PaulKnight> vote: 3,4a
<pauld> chad, count
<chad> Question: options for CR18 and CR20
<chad> Option 3: Paco's option 3: (8)
<chad> Option 3a: Paco's option 3 but for SOAP 1.1 only. (0)
<chad> Option 4: Use SOAP 1.2 ImmediateDestination property (3)
<chad> Option 4a: make [destionation] optional and remove defaulting to anon (5)
<chad> 16 voters: agupta (4a) , anish (4a) , David_Illsley (3) , dhull (4, 3a) , Gil (4a, 4, 3a) , GlenD (4a, 4, 3a) , hugo (4, 4a, 3, 3a) , Jonathan (3) , Katy (3) , Paco (3) , pauld (3) , PaulKnight (3, 4a) , TonyR (4a, 3) , TRutt (3, 4a) , uyalcina (3) , vikas (4)
<chad> Round 1: Count of first place rankings.
<chad> Round 2: First elimination round.
<chad> Eliminating candidadates without any votes.
<Nilo> 4a, 3a, 3
<chad> Eliminating candidate 3a.
<chad> Round 3: Eliminating candidate 4.
<chad> Round 4: Eliminating candidate 4a.
<chad> Candidate 3 is elected.
<chad> Winner is option 3 - Paco's option 3:
<Nilo> vote: 4a, 3a, 3
<anish> vote: nilo: 4a, 3a, 3
<pauld> chad, count
<chad> Question: options for CR18 and CR20
<chad> Option 3: Paco's option 3: (8)
<chad> Option 3a: Paco's option 3 but for SOAP 1.1 only. (0)
<chad> Option 4: Use SOAP 1.2 ImmediateDestination property (3)
<chad> Option 4a: make [destionation] optional and remove defaulting to anon (7)
<chad> 18 voters: agupta (4a) , anish (4a) , David_Illsley (3) , dhull (4, 3a) , Gil (4a, 4, 3a) , GlenD (4a, 4, 3a) , hugo (4, 4a, 3, 3a) , Jonathan (3) , Katy (3) , Nilo (4a, 3a, 3) , nilo (4a, 3a, 3) , Paco (3) , pauld (3) , PaulKnight (3, 4a) , TonyR (4a, 3) , TRutt (3, 4a) , uyalcina (3) , vikas (4)
<chad> Round 1: Count of first place rankings.
<chad> Round 2: First elimination round.
<chad> Eliminating candidadates without any votes.
<chad> Eliminating candidate 3a.
<chad> Round 3: Eliminating candidate 4.
<chad> Round 4: Tie when choosing candidate to eliminate.
<chad> Tie at round 3 between 3, 4a.
<chad> Candidate 4a has the fewest votes at round 2.
<chad> Eliminating candidate 4a.
<chad> Candidate 3 is elected.
<chad> Winner is option 3 - Paco's option 3:
<anish> chad, details?
<pauld> vote: Nilo: abstain
<uyalcina> +1 to JM
<pauld> want's to clarify, there are no longer any anonymous, or missing To's in the test suite
<GlenD> I think 4a solves the problem in a better way, and I'd prefer that <To> still be allowed to be missing. Not all implementations need deal with that, of course.
<GlenD> We don't actually have to say what anonymous means in <To> at all, just like we don't say what foo:something/or/other means as a <To>
<uyalcina> i disagree Glen. With 4a, you still need to allow/disallow wsa:To anonymous for a request message for SOAP/HTTP
<GlenD> Why?
<uyalcina> you have not solved the problem at all with 4a
<GlenD> Anonymous has special meaning for a response EPR (ReplyTo, FaultTo). Outside that context, we don't have to say anything, just like we don't say anything in particular about OTHER URIs.
<GlenD> i.e. if your implementation asks for a <To> with a particular URI, it's up to you to understand it.
<uyalcina> i am sorry, undefined stuff do not do it for me. The reason we have cr18 is because we have an undefined situation
<scribe> ACTION: Anish to put 4a into "crisp text" [recorded in http://www.w3.org/2006/02/06-ws-addr-minutes.html#action01]
<TRutt> yes
Bob: intent is to clarify differences between 3 and 4a and vote next week
<pauld> dispatching can happen on whatever you want
<TRutt> Some implementations do not make the http post url available at the soap level
<pauld> action, GED, if it's raining, which advert you prefered in the superbowl .. whatever
<GlenD> +1 pauld
<anish> i agree too that dispatching can happen on whatever you want
Bob: last week in April or first week in May
Tom Rutt: would prefer first week in May
Bob: For planning purposes then, the face to face after Cannes will be week of May 1 with a preference heard for May 3 and 4. Probable location is the East Coast
<anish> oasis symp. is in San Fran
ADJOURN