Re: New Alternative F to resolve LC comment on WS addr metadata

Tom Rutt wrote:

I forgot to delete the following paragraph in the last section of the 
proposed changes.
"
Consider the following example, where we have a client who does not care
whether the endpoint explicitly supports anonymous responses, and a WSDL
which states that the endpoint does explicitly support anonymous responses.
"

Since this change is so simple, I have not redistributed the pdf with 
this new change.

Tom
> During off line discussions during the ws-policy meeting, a group of 
> people came up with
> another alternative to resolve the LC comment on WS addr metadata.
>
> The most significant concern about the existing nested support 
> assertions is that there
> is no way to indicate that a particular response mode is not supported.
>
> This new alternative F takes the approach of nested support 
> assertions, however
> non presence of a nested policy assertion now implies that the 
> associated response mode is not supported.
>
> An empty Addressing assertion would imply that addressing is required, 
> but that no response
> modes associated with EPRs in a request message are supported.
>
> It is important to notice that the use of negative language 
> (incljuding the word "Negation" ) is never used in the definitions for 
> Alternative F.
>
> I deleted the example which places the two nested assertions in 
> separate alternatives, since
> such a policy expression is not useful for any endpoint.
>
> I really like this new alternative, since it allows expression of all 
> support claims an endpoint would  ever want to assert, and also allows 
> for client "expressions" to allow use of intersection to find  
> compatible endpoints.
>
> Tom Rutt
>
> I really like this new alternative.
> of the new assertions.
>


-- 
----------------------------------------------------
Tom Rutt	email: tom@coastin.com; trutt@us.fujitsu.com
Tel: +1 732 801 5744          Fax: +1 732 774 5133

Received on Thursday, 15 March 2007 19:38:23 UTC