Web Services Description -- 29 Jun 2006

Tony: Since we declined CR044, we will include this resolution as a suggestion for language to clarify the rationale for default properties in support of interface-less bindings

Jacek: We need to correct the language suggested for section 6.8 as per http://lists.w3.org/Archives/Public/www-ws-desc/2006Jun/0026.html

ACTION: Charlton to update his proposal for CR044 based on the language correction suggested by Jacek per http://lists.w3.org/Archives/Public/www-ws-desc/2006Jun/0026.html [recorded in http://www.w3.org/2006/06/29-ws-desc-minutes.html#action04]

CR045

http://www.w3.org/2002/ws/desc/5/cr-issues/#CR045

Unfortunately Paul just dropped off the telcon, so we'll postpone discussion of this issue

CR052

http://www.w3.org/2002/ws/desc/5/cr-issues/#CR052

Serializing only part of the body with HTTP binding

(non-sequitur to CR044 - no objections to accepting the proposal; will forward it to the editors as described)

CR052 wants to speak toward updating the state through an append

Roberto: Not a trivial change

Jacek: Not serializing the location as part of the payload

Philippe: This is a new use case. A bit late in the cycle to include it

Tony: Input for the next revision

Jacek: Or toward producing a new extension

Tony: e.g. HTTP 2.0 binding. Difficult to handle a new use case at this time, agreed. We'll need to respond likewise.

Jacek volunteers to address this.

ACTION: Jacek to draft a response to Eric re: CR052 explaining that this represents a new use case, and that we will not be able to address this as such in the spec, but that it can be addressed as an additional extension... [recorded in http://www.w3.org/2006/06/29-ws-desc-minutes.html#action01]

CR053

http://www.w3.org/2002/ws/desc/5/cr-issues/#CR053

Allow absolute URI in {location}

Jacek: I don't think it is possible to craft the URI inclusive of extended variables
... I think what he wants is that the location be considered for combination with the base IRI

Philippe: Don't see anything that makes that impossible

<JacekK> location="http://foo"; address="http://example.com/"; whttp:location="{location}"

<JacekK> http://example.com/{location}

<JacekK> http://example.com/http://foo

<Roberto> 6.4.2. says:

<Roberto> Input serializations may define additional processing rules to be applied to the value of {http location} before combining it with the {address} property of the endpoint element to form the HTTP request IRI.

<Roberto> http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl20/wsdl20-adjuncts.html?content-type=text/html;%20charset=utf-8#http-operation-decl-relate

<gil> it seems like there are two issues here

Jacek: Need to clarify and provide an example - to illustrate where the address will be ignored and not included in the path - it isn't obvious

<Roberto> 6.4.2: This IRI is combined with the base IRI specified in the {address} property of the Endpoint component to form the full IRI for the HTTP request to invoke the operation.

Roberto: Also need to clarify what is meant as the combination of IRIs

<gil> (a) seems to be another new use case around "following a hyperlink"

<Roberto> where are the combination rules?

<gil> (b) seems to be clarifying the spec

Jacek: I think the spec supports this as Roberto described. Only need to clarify that it is in fact true, and that it may have an unexpected effect in certain corner cases

Philippe: Are we doing some URI escaping or not?

Tony: Only response is that there is no escape. What Roberto appears to have pointed out is that it substitutes the location before deciding what to do with what amounts to the two URIs

Jacek: No escaping mentioned in 6.7

Tony: If provide substitution by itself, implying "use URI as URI, don't escape it", but if in a string, then it is a parameter

Jacek: If anything before the parameter, escape it, otherwise not

Roberto: Way to address this would be to add more syntax

Philippe: Add explanatory text to spec - no encoding at this stage, and encoding only once full URI is formed. We would need to add a warning re: security w.r.t. this. Someone can make your service to go somewhere else if you don't check what you use as input.

Tony: So we should clearly indicate that no escaping is going on?

Jacek: Someone can interpret our spec to indicate that escaping is going on

Philippe: We need to clarify the spec language around this

ACTION: Philippe to write up recommended text to clarify the issue in CR53 [recorded in http://www.w3.org/2006/06/29-ws-desc-minutes.html#action02]

CR054

Tony: Sounds like Eric is asking for a change to XForms

Jacek: I think we should correct Eric on this. Specifies element name, not attribute, in section 6.1.1

Philippe: Don't believe there is an issue here. Eric is providing feedback as requested.

ACTION: Jacek to write a response to Eric correcting his interpretation of the text as described in CR054 [recorded in http://www.w3.org/2006/06/29-ws-desc-minutes.html#action03]

CR055

Jacek: This is done

http://www.w3.org/2002/ws/desc/5/cr-issues/#CR055

http://lists.w3.org/Archives/Public/www-ws-desc/2006May/0087.html

<JacekK> Jacek's action completed at http://lists.w3.org/Archives/Public/www-ws-desc/2006Jun/0026.html

Clarification needed on HTTP Transfer Coding

<JacekK> http://lists.w3.org/Archives/Public/www-ws-desc/2006Jun/0032.html

I believe the desired description is clearly made with the language at hand with the proposal for CR044

Tony: If we accept Charlton's language for CR044 than we can consider this issue closed as well

http://lists.w3.org/Archives/Public/www-ws-desc/2006Jun/0073.html

Tony: CR055 to be left open pending Charlton's AI for CR044

CR056

http://www.w3.org/2002/ws/desc/5/cr-issues/#CR056

Questions on {http method} and {safety} extension properties

Tony: Mark CR056 as pending based on administrivia 4c

[Philippe fixed the reported typo]

CR057

http://www.w3.org/2002/ws/desc/5/cr-issues/#CR057

HTTPHeader name should be xs:string, not xs:QName

Jacek: Does anyone disdisagree with John Kaputin on this?

Gil: Note that John has implemented HTTPHeader as xs:string in Woden

Jacek: Have an example...

Jacek reads proposal

Resolution: Accept John's proposal as editorial item

ACTION: Editors to ensure that HTTPHeader name is of type xs:string, and not XS:Qname (see issue CR057)

Tony: At the end of time for the telcon - see you on the call next week

Web Services Description
29 Jun 2006

Attendees

Contents

Approval of minutes

Review of action items

Administrivia

CR022

CR037

CR044

CR045

CR052

CR053

CR054

CR055

CR056

CR057

Summary of Action Items

Web Services Description 29 Jun 2006

Attendees

Contents

Approval of minutes

Review of action items

Administrivia

CR022

CR037

CR044

CR045

CR052

CR053

CR054

CR055

CR056

CR057

Summary of Action Items

Web Services Description
29 Jun 2006