W3C home > Mailing lists > Public > www-ws-desc@w3.org > February 2003

Re: Action 2003-01-21 for Umit

From: Sanjiva Weerawarana <sanjiva@watson.ibm.com>
Date: Thu, 6 Feb 2003 08:33:48 +0600
Message-ID: <029501c2cd88$322d9690$b100a8c0@lankabook2>
To: "Martin Gudgin" <mgudgin@microsoft.com>, "Umit Yalcinalp" <umit.yalcinalp@oracle.com>, "Don Mullen" <donmullen@tibco.com>
Cc: <www-ws-desc@w3.org>

"Martin Gudgin" <mgudgin@microsoft.com> writes:
>   "How does your mapping scheme account for mapping parts to headers?
> Are they accounted for in the schema as well?"
> is that we decided in Virginia that SOAP headers are declared in a
> binding by direct reference to either a global element declaration or a
> complex type definition. In the latter case, a local name and namespace
> name are provided in the binding to specify the [localname] and
> [namespace name] properties of the header EII. So the status quo is that
> SOAP headers are NEVER parts of a WSDL message.

As I said at the Virginia F2F, this is wrong because this approach
only works for SOAP. This significantly de-values the abstract description
contained in WSDL as a key part of the exchange, that of headers, is only 
found within the SOAP  binding section.

The context proposal I sent is a proposal for how to fix this problem.
The header items would then be defined (still as XML Schema types) as
contextual items. Then, the operations of the portTypes would be 
annotated to show how these values are to be exchanged. Finally, a 
SOAP binding would indicate how they flow as headers. 

The result is that the header data are not in the contents of <message>
(or its replacement). However, they too are described abstractly first
to enable one to bind the information to another protocol. IMHO not 
doing this will basically mean we've made the entire portType stuff 
totally irrelevant as describing information that will eventually flow
as headers are a critical part of documenting a service.

Jonathan, I'd like to get the context proposal on a telecon agenda 
soon please.

Received on Wednesday, 5 February 2003 21:36:48 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:54:41 UTC