See also: IRC log
APPROVED: April 22nd and April 29th minutes
<scribe> ACTION: Pauld to write up examples of schemas for the Primer (2004-01-08) [DONE]
<scribe> ACTION: Philippe and JMarsh will look at the ipr for test suite (2004-01-28) [PENDING]
<scribe> ACTION: DaveO to produce a refined proposal for Asynch HTTP binding addressing the concerns of folks that object to leaving replyTo info out of WSDL (2004-02-12) [PENDING]
<scribe> ACTION: Marsh will get schema tf going (2004-04-01) [PENDING]
<scribe> ACTION: Arthur to write up concerns about XML 1.1 implications on WSDL 2.0 for potential forwarding to XML WG (2004-04-29) [PENDING]
<scribe> ACTION: Hugo to redraft DaveO's proposal for extensible HTTP operations (2004-04-29) [DONE]
<scribe> ACTION: Marsh to add an issue on describing the XML version of SOAP messages (2004-04-29) [DONE]
<scribe> ACTION: Umit to work with Anish to create a working draft on media types ready for our May F2F meeting in NYC (2004-04-29) [PENDING]
<scribe> ACTION: Part 1 editors to adopt Jacek's "purpose of the binding" text, without "interchangeable"endpoints, and using "confidentiality" (or similar) instead of TLS (2004-04-29) [PENDING]
Ugo Corda replaces Alan for See Beyond
Latest proposal from Hugo (DaveO+Hugo's combined proposal), http://lists.w3.org/Archives/Public/www-ws-desc/2004Apr/0093.html
marsh: should we walk thru from the top or discuss any one item in specific?
...is webdav useful to support?
youenn: how far do we support http features for webdav?
Yaron: how far do we support http features for webdav?
marsh: there are things that can be discovered @runtime, not sure how much of these would be useful in a wsdl description
yaron: qname does not allow you to disambiguate, specifying redirect ends up being useful,
glen: address uri is the only thing that is used to disambiguate endpoints
yaron: may be using other pieces such as ws-addressing, don't want to go there tho'
<Marsh> binding="xs:QName" >
marsh: why would you like to know about re-direction via wsdl?
glen: what if my implementation does not support re-direction
dbooth: there are cases where you do not want to be re-directed
glen: I would say a ? on re-direction, not really useful
...status code and response status
marsh: anything not useful should not be considered .. lets consider useful vs. !useful
HTTP Version (DaveO=probably useful, glen=useful
Content coding (DaveO=probably useful, glen=useful
glen: may depend on the context
yaron: yes, they may change across time .. then its got to go thru the http negotiation .. cost is one round trip
marsh: but, what about the contract?
...what level are we describing?
...level 1 - that can go thru negotiation, level 2 for optimization, are there only 2 buckets
...what does it mean to describe in wsdl?
pauld: wsdl is a contract. If you change it, then you broke it
yaron: if you make a promise via wsdl, then don't violate it .. if you do then http offers a fall back
pauld: some of the items are useful to default, some of the items are required ..
dbooth: in the case of gzip, perhaps the contract should just say that it supports a certain version of HTTP, then it could withdraw support for gzip without violating its contract
glen: contract, but this is separate from the runtime message .. to some extent an implementation should be able to deal with these changes. it is useful if you do not violate the contract.
marsh: useful vs. !useful, optimization (hint), required features
HTTP Version (DaveO=probably useful glen=says 1.1 then 1.0 should be ok
dbooth: disagrees, that is a violation
HTTP Version in the optimization bucket
marsh: Transfer Codings (Chunked encoding) in the optimization bucket?
...summarizes that chunked encoding goes in the optimization bucket
<dbooth> If a service declares in its WSDL doc that it supports HTTP 1.1, then it must support HTTP 1.1. However, the client has the prerogative of speaking HTTP 1.0 to it, because HTTP 1.1 is backwards compatible with HTTP 1.0.
Authentication (DaveO=probably useful
...cookies, required .. if you don't support this then you cannot interact with me
...optimization in which parts of the message, should be able to declare it
moving thru the list
yaron: if you are doing caching, here are some useful things ..
...you may want to have a caching module to use it
marsh: caching is in the optional bucket
Content Negotiation (DaveO=probably useful
reading from daveo's e-mail .. probably useful in the MTOM context
marsh: sounds like optimization
yaron: in theory no .. a little on the bizarre side
marsh: how do I use it in the wsdl context
yaron: we should ask MTOM folks if they want to describe it at the http level
...there are lots of ways to serialize xml .. this is a way to declare a way to declare that support .. say, i am going to mark it in http
marsh: looks like an optimization, lets you to specify a default .. but, you are saying that this may be spelling out a requirement
...this one can go either way
Cookies (DaveO=probably useful
marsh: summarizes Arthur's position
...required=authentication, cookies, ssl - are there positions on whether we should describe them
(no answers ..)
jeff: can't find a reason why we should not
marsh: these are requirements for a client to talk with a server
...optimization bucket .. onto e-mail
...requirements & capabilities bucket .. soliciting volunteers
...we are looking for volunteers
<scribe> ACTION: Marsh to draft David O or Hugo to write up a proposal for describing the requirement & capabilities bucket of http features (2004-05-06)
marsh: lets continue to review the list again and classify them
marsh: any comments on JJM's proposed schema
yaron: message may have multiple media types, it is not clear ?
glen: not in SOAP 1.2 tho'
marsh: shall we remove media type from this proposal
jjm: have no problems with it
<pauld> Sanjiva + Roberto's alterations to the fault code descriptions ?
jjm: may be .. remove wsoap:header, perhaps should not appear in WSDL
<Zakim> Roberto, you wanted to comment on wsoap:header
roberto: comment on wsoap:header; don't quite understand why declare mustUnderstand, role,
glen: mustUnderstand is little weird
...you are saying .. target it at this role
yaron: app may depend on specific processing path, example logging, .. role is useful
jjm: mustUnderstand in conjunction with a role
glen: walks thru an example, cache control intermediary
...more details ..
roberto: not too convinced, can live with it
(sliding into intermediary discussion ..)
glen: may designate parts of the proposal as pending further intermediary discussion
roberto: take them out until we discuss intermediary
glen: features and properties are relavant to this discussion
marsh: schema will continue to have modules; ok for the editors to add them?
marsh: happy to incorporate this proposal, remove media type, parts that are relevant to intermediary (mustUnderstand and relay)
<scribe> ACTION: JJM to incorporate this proposal (SOAP 1.2 Binding) after removing parts that are relevant to media type and intermediary (2004-05-06)
Thread, http://lists.w3.org/Archives/Public/www-ws-desc/2004Apr/0081.html (sanjiva)
Roberto's amendment, http://lists.w3.org/Archives/Public/www-ws-desc/2004Apr/0081.html
yaron: likes sanjiva's proposal
asir: likes sanjiva's proposal, intuitive ..
marsh: lets take Roberto's amendment
roberto: willing to withdraw if there is no support
...lets go with sanjiva proposal and ..
DECISION: adopt Sanjiva's proposal
issue 172 is closed
<dbooth> NYC F2F schedule: http://www.w3.org/2002/ws/desc/4/04-05-f2f.htm#Schedule
glen: talks about the F2F goals
marsh: close all of the issues in part 3
...once jjm's proposal is incorporated, then we may have more issues from the WG
...if we get thru part 3 issues, then we will go to part 1 issues, test deliverables, ..
...hope we have a clean part 3 draft for the wg to review
dbooth: we have a new charter that spells out schedule as the priority
yaron: we have made great progress ..