Re: Action item - lc129

Hi Marc,

I am not sure why you say it is not testable. If the <wsaw:Anonymous> 
element is not present then you need to show me metadata indicating what 
to do about the anonymous URI - a WSDL binding, a policy etc. If you do 
(say, WCF comes up with a WSDL binding) you are ok, otherwise you are not 
(and need to give a good reason why not). So I think using SHOULD does 
make sense. 

A separate issue is what kind of metadata is "acceptable", but I would 
claim that it is not in the WS-Addressing spec interest to define a close 
set of possible metadata formats that might be used for this purpose. At 
the interop test we may need to settle for one or another, but in general 
we don't want to write the spec so narrowly that expected future use cases 
fall completely outside its scope. 

Paco




Marc Hadley <Marc.Hadley@Sun.COM> 
Sent by: Marc.Hadley@Sun.COM
04/24/2006 03:30 PM

To
Francisco Curbera/Watson/IBM@IBMUS
cc
"[WS-A]" <public-ws-addressing@w3.org>
Subject
Re: Action item - lc129






On Apr 16, 2006, at 10:31 PM, Francisco Curbera wrote:
>
> This is my take on expanding "option 4" in Jonathan's mail [1] 
> ("Remove the
> default. Lack of wsaw:Anonymous means there are no claims about 
> Anonymous
> support."). I am not proposing here the changes necessary to fully
> incorporate a resolution of the issue, only proposing a 
> clarification of
> the assumptions clients would be able to make when no wsaw:Anonymous
> element is present.
>
> "A WSDL or policy based service description that includes the
> wsaw:UsingAddressing but no a wsaw:Anonymous marker makes no assertion
> regarding a requirement or a constraint in the use of the anonymous 
> URI in
> EPRs contained in messages sent to the endpoint. In this cases, 
> endpoint
> service descriptions SHOULD use additional metadata, such as WSDL 
> bindings
> or additional policy assertions, to indicate any requirements or
> restrictions on the use of the anonymous URI by clients.

I'm not sure about the SHOULD in the above. I'd only expect an RFC 
2119 keyword if we were recommending doing something specific that 
could help with interop. I don't think it makes sense to recommend 
(in an RFC 2119 sense) doing something without specifying exactly 
what since this is an inherently untestable assertion.

Marc.

> However, in the
> absence of additional metadata, clients of the endpoint MAY assume 
> that the
> service endpoint follows the behavior indicated by the 'optional' 
> value of
> the wsaw:Anonymous marker. An endpoint MAY send a fault back to the 
> client
> if a message received uses the anonymous URI in a way that is 
> unsupported
> by the endpoint."
>
> [1].
> http://lists.w3.org/Archives/Public/public-ws-addressing/2006Apr/ 
> 0019.html
>
>

---
Marc Hadley <marc.hadley at sun.com>
Business Alliances, CTO Office, Sun Microsystems.