RE: i022: Relationship to the SOAP Binding Framework: a SOAP Request MEP and OneWay HTTP Binding

I disagreed with your assessment - and ultimately the majority of the
WSDL wg - then and I disagree now.  IMO, the SOAP WG did not say that a
WSDL in had to be in the HTTP request and the WSDL out had to be in the
HTTP response for a WSDL in-out using the SOAP HTTP binding.  But it
doesn't really matter whether it's the WSDL WG or the WSDL WGs
interpretation of the XMLP WGs response, it's not permitted in WSDL 2.0

 

I think this is the heart of this particular bit of difficulty: that
SOAP is both a format and a protocol.  And it's describing the protocol
parts that is causing us this heartburn.  

 

Dave

 

  _____  

From: Yalcinalp, Umit [mailto:umit.yalcinalp@sap.com] 
Sent: Thursday, December 23, 2004 3:23 PM
To: David Orchard; Jonathan Marsh; public-ws-addressing@w3.org
Subject: RE: i022: Relationship to the SOAP Binding Framework: a SOAP
Request MEP and OneWay HTTP Binding

 

Well, to be more accurate, WSD wg has been discussing (and posted
requests for clarifications from the XMLP wg) the differentiation
between the cases 3 and 4. The question has been whether the SOAP
req-response binding actually allows the callback scenerio at all,
namely #4. Therefore, it is not really the WSD wg's enforcement per se.
It is the definition of SOAP request-response MEP that WSD wg has to
work with. 

 

--umit

 

	-----Original Message-----
	From: public-ws-addressing-request@w3.org
[mailto:public-ws-addressing-request@w3.org] 
	Sent: Thursday, Dec 23, 2004 14:47 PM
	To: Jonathan Marsh; public-ws-addressing@w3.org
	Subject: RE: i022: Relationship to the SOAP Binding Framework: a
SOAP Request MEP and OneWay HTTP Binding

	Good question.  The WSDL SOAP binding would specify the one-way
MEP for a wsdl in-out to tell the binding whether it's a soap
req-response binding or 2 soap one-way bindings.   I don't think it is
that useful for a wsdl in only mep to distinguish between request vs
request response mep as the one-way or the req-resp binding seem to be
the same thing in the one way case.

	 

	Some of the possible scenarios I see are:

	1. wsdl in, soap request mep, soap one-way binding

	2. wsdl in, soap request-resp mep, soap req resp binding

	3. wsdl in-out, soap req-resp mep, soap req resp binding

	4. wsdl in-out, soap request mep, 2 soap one-way bindings (the
callback scenario)

	5. wsdl in-out, soap req-resp mep, 2 soap req resp bindings

	 

	Note that #5 - which I proposed - has been precluded by the WSDL
2.0 WG.  Effectively, the WSDL 2.0 group forced the creation the request
mep for wsdl in-outs that is a callback over 2 http connections.

	 

	Cheers,

	Dave

	 

	
  _____  


	From: Jonathan Marsh [mailto:jmarsh@microsoft.com] 
	Sent: Thursday, December 23, 2004 1:59 PM
	To: David Orchard; public-ws-addressing@w3.org
	Subject: RE: i022: Relationship to the SOAP Binding Framework: a
SOAP Request MEP and OneWay HTTP Binding

	 

	How precisely would one use this MEP in the WSDL SOAP binding?
Specifically I'm wondering if there will ever be cases where it's not
clear whether to use this MEP or the existing SOAP Request-Response MEP.

	 

	
  _____  


	From: public-ws-addressing-request@w3.org
[mailto:public-ws-addressing-request@w3.org] On Behalf Of David Orchard
	Sent: Friday, December 17, 2004 2:06 PM
	To: public-ws-addressing@w3.org
	Subject: i022: Relationship to the SOAP Binding Framework: a
SOAP Request MEP and OneWay HTTP Binding

	 

	I offer up the first new W3C SOAP MEP and Binding in many a
moon...

	 

	I've written up a first draft of a proposed WS-Addressing
Adjuncts with a SOAP request MEP and a one-way SOAP HTTP Binding.  I
believe this is sufficient to close issue 22.  

	 

	I think it allows the soap request-response MEP to be layered on
2 one-way SOAP HTTP Bindings, but I haven't really verified it.  I've
had a few problems with the links, which I will work on once y'all have
had a chance to review.  I basically copied the soap 1.2 adjuncts MEP
and Binding section.

	 

	Some of the tricky areas that I thought I'd call out:

	- There is an optional binding specific response in the one-way
MEP.

	- Relationship to media type.  I think this be covered by the
soap media-type, but I'm not 100% sure.  I'm not sure about the case of
whether a soap+xml is good enough for this mep+binding, I sure hope so
though.

	- Webmethod support: I said POST only

	- The identification of the mep in use can't be gleaned from the
information in the binding, unlike the SOAP HTTP Binding

	- Streaming: I consistently said that requesting SOAP nodes must
avoid deadlock by accepting binding-specific response messages

	- I removed "receiving" state from the next state tables.

	- SOAP faults cannot come back over the http response.  For
request-response bound to 2 http requests, life sucks.

	- The binding can allow an empty body, especially for cases
where the action is sufficient.

	- I kept the HTTP status code at 200

	 

	I would also like to mention that I found this exercise very
informative.  I think that SOAP has provided an excellent framework for
creating interoperable meps and bindings as it forced me to think about
many hard issues.

	 

	Cheers,

	Dave