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

Re: Corrected Proposed Alternative A for resolution of ws addr metadata LC comment

From: Katy Warr <katy_warr@uk.ibm.com>
Date: Mon, 5 Mar 2007 13:55:15 +0000
To: tom@coastin.com
Cc: WS-Addressing <public-ws-addressing@w3.org>, ws policy <public-ws-policy@w3.org>
Message-ID: <OFD86FCD1C.9534DF0A-ON80257295.00490C75-80257295.004C7906@uk.ibm.com>
Tom,

This sentence implies a requirement on the request message:

"The appearance of this element within a policy alternative indicates that 
the
subject requires any request message that has responses to include 
response
endpoint EPRs that contain the anonymous URI
("http://www.w3.org/2005/08/addressing/anonymous") as the value of 
[address]."

Whereas, this sentence implies that the requirement is scoped to the 
response message:

"In other words, the subject requires that response instances are sent 
using the
anonymous URI."

I personally think that explaining the requirement in terms of what is 
required by the client request message is preferable because the whole 
point of this service policy is for client<->service policy negotiations. 
The client needs to narrow the available service policies to a single 
alternative prior to sending the request. 

If we agreed that nested policies apply to client requests (rather than 
responses), we can still fix the problem of  mixing replyTo-anon and 
faultTo=nonAnon.  As the proposal stands, 
AnonymousResponse/NonAnonymousResponse contradict each other so cannot 
exist together within a single alternative. This could be fixed by:
1) adding a 3rd nested assertion indicating that the request message must 
contain differing values for replyTo and faultTo (one anon, the other 
non-anon).  Eg 'AnonymousAndNonAnonymousResponses'.  This has the benefit 
of allowing service providers to omit this assertion if they were unable 
to support this more complex mixed-response type scenario. 
2) alternatively, we could revert back to specifying support rather than 
requirement thus allowing the nested assertions to co-exist within a 
single alternative.  This has the benefit that it makes the nested 
assertions composable, but it goes against the direction indicated by the 
wsp team.

thanks,
Katy




Tom Rutt <tom@coastin.com> 
Sent by: public-ws-addressing-request@w3.org
04/03/2007 22:20
Please respond to
tom@coastin.com


To
Katy Warr/UK/IBM@IBMGB
cc
WS-Addressing <public-ws-addressing@w3.org>, ws policy 
<public-ws-policy@w3.org>
Subject
Re: Corrected Proposed Alternative A for resolution of ws addr  metadata 
LC   comment







Well that case can still be handled by having two or more nested 
alternatitves.


Remember the alternatives are for responses not requests.  Each reply to 
and fault to
is a separate response.

However, I still in more in favor of alternative D) to delete the nested 
policy assertions
for Addressing assertion


Tom

Katy Warr wrote:
>
> Tom
> We have discussed scenarios in previous wsa meetings where a client 
> may need to mix anon/non-anon on a single request:
> - e.g.  replyTo=anon, faultTo=nonAnon
> As wsam:AnonymousResponses and wsam:NonAnonymousResponses are 
> contradictory requirements, there does not appear to be a way for a 
> service indicate that it could accept this combination on a single 
> request (as these assertions specify a single hard requirement which 
> applies to both replyTo and faultTo).
> regards,
> Katy
>
>
>
> *Tom Rutt <tom@coastin.com>*
> Sent by: public-ws-policy-request@w3.org
>
> 02/03/2007 20:36
> Please respond to
> tom@coastin.com
>
>
> 
> To
>                WS-Addressing <public-ws-addressing@w3.org>, ws policy 
> <public-ws-policy@w3.org>
> cc
> 
> Subject
>                Corrected Proposed Alternative A for resolution of ws 
addr metadata 
> LC  comment
>
>
>
> 
>
>
>
>
>
> I never posted a complete change proposal for Alternative A (which
> defines the nested assertions of Addressing assertion as requirements,
> with absence implying Prohibition.
>
> I post this for completeness.
>
> -- 
> ----------------------------------------------------
> Tom Rutt                 email: tom@coastin.com; trutt@us.fujitsu.com
> Tel: +1 732 801 5744          Fax: +1 732 774 5133
>
>
> [attachment "ws-AddrMetadataPolicyEdits-altA.pdf" deleted by Katy 
> Warr/UK/IBM]
>
>
>
> ------------------------------------------------------------------------
>
> /
> /
>
> /Unless stated otherwise above:
> IBM United Kingdom Limited - Registered in England and Wales with 
> number 741598.
> Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 
> 3AU/
>
>
>
>
>
>

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










Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
Received on Monday, 5 March 2007 13:56:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:35:16 GMT