W3C home > Mailing lists > Public > public-ws-addressing@w3.org > January 2007

RE: retaining the {addressing} property?

From: Jonathan Marsh <jonathan@wso2.com>
Date: Wed, 10 Jan 2007 08:20:50 -0800
To: "'David Illsley'" <david.illsley@uk.ibm.com>, "'Rogers, Tony'" <Tony.Rogers@ca.com>
Cc: <public-ws-addressing@w3.org>
Message-ID: <013101c734d3$4e4e3300$3501a8c0@DELLICIOUS>

IMO, it would be weird to have certain assertions (WS-Addressing) result in
direct WSDL component model changes, and others not (security policy for
instance).  If UsingAddressing goes, so should {addressing} I think.

Jonathan Marsh - http://www.wso2.com - http://auburnmarshes.spaces.live.com
 
> -----Original Message-----
> From: David Illsley [mailto:david.illsley@uk.ibm.com]
> Sent: Wednesday, January 10, 2007 5:37 AM
> To: Rogers, Tony
> Cc: public-ws-addressing@w3.org; Jonathan Marsh
> Subject: Re: retaining the {addressing} property?
> 
> I'm afraid my WSDL 2.0 and Policy Attachment for WSDL 2.0 knowledge isn't
> up to much.
> 
> It appears from WS-Policy Attachment [1] that the Addressing assertion
> will appear as part of the {policy} property of the relevant WSDL
> component. That may mean that we can reference that property to define
> where to look for the assertion... but the fact that the value then
> appears twice in component model and only once in the XML seems strange.
> My impression of WSDL 2.0 is that it's expected that an extensibility
> property maps to an extensibility element or attribute.
> 
> This makes me wonder if we should be removing the {addressing} WSDL 2.0
> component model property. Can any WSDL 2.0 experts out there comment on
> this? Jonathan?
> 
> Sorry Tony, not an answer to the question you asked,
> David
> 
> [1]
> http://www.w3.org/TR/2006/WD-ws-policy-attach-20061117/#ws-policy-
> attachment-for-wsdl20
> 
> David Illsley
> Web Services Development
> MP211, IBM Hursley Park, SO21 2JN
> +44 (0)1962 815049 (Int. 245049)
> david.illsley@uk.ibm.com
> 
> 
> 
> From:
> "Rogers, Tony" <Tony.Rogers@ca.com>
> To:
> <public-ws-addressing@w3.org>
> Cc:
> David Illsley/UK/IBM@IBMGB
> Date:
> 01/10/2007 10:48 AM
> Subject:
> retaining the {addressing} property?
> 
> 
> We are removing the UsingAddressing WSDL marker. I naively assumed that
> meant removing the entire section 3.1. Upon more detailed study, I am
> fairly sure that that is NOT the case.
> 
> Do we intend to retain the {addressing} property which 3.1.1 introduced as
> an extension to the WSDL 2.0 component model? I suspect that we need to do
> so.
> 
> Should we describe how the presence / absence / wsp:Optional-ness of the
> wsam:Addressing assertion affects the {addressing} property? It can be
> done, but I'd appreciate some help in phrasing it. Perhaps some words
> along the line of:
> 
> The {addressing} property is present if the wsam:Addressing assertion is
> present. The value of the property is "required" is the only policy
> alternatives present include this assertion, without the
> wsp:Optional="true" attribute . The value of the property is "optional" if
> there are policy alternatives that include this assertion, and
> alternatives which do not include it, or if it is marked with the
> wsp:Optional="true" attribute.
> 
> I'd really appreciate a Policy-literate person rephrasing that in
> legitimate Policy language :-) Should I separate the discussion of
> alternatives with and without the assertion from discussion of
> wsp:Optional?
> 
> Do we also retain Table 3-1 showing the effect of the {addressing}
> property? I think this table, plus the discusson of what must be present
> for the message to be compliant, belongs in a separate section, placed
> after the discussion of how the property is set. The current layout
> confuses things, because it mixes the defining of the property with the
> effects of the property, thus clouding the discussion of the SOAP module's
> ability to affect the same property.
> 
> So I think we should have 3.1 specifying the policy assertions, and their
> effect on the {addressing} property, then 3.2 specifying the SOAP module,
> and its effect on the {addressing} property, then 3.3 describing the
> {addressing} property and the presence/absence of MAPs in the message.
> 
> Does that make sense?
> 
> Tony Rogers
> CA, Inc
> Senior Architect, Development
> tony.rogers@ca.com
> 
Received on Wednesday, 10 January 2007 16:40:37 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:35:15 GMT