RE: Issue i018 -- EPR abstract properties and binding to SOAP

Why wouldn't the EPR minter just put both sets of attributes on the Ref
Params? e.g.

<m:MyRefParam soap11:actor='foobar' soap12:role='foobar' />

Gudge 

> -----Original Message-----
> From: public-ws-addressing-request@w3.org 
> [mailto:public-ws-addressing-request@w3.org] On Behalf Of 
> Anish Karmarkar
> Sent: 07 February 2005 08:25
> To: public-ws-addressing@w3.org
> Subject: Issue i018 -- EPR abstract properties and binding to SOAP
> 
> 
> I took an action to identify which parts of the specification give an 
> impression that EPRs are more abstract than they actually are. This 
> email fulfills that action.
> 
> This action was assigned during the discussion of issue i018 
> at the MEL 
> f2f. This issue deals with the fact that the SOAP binding for 
> EPR maps 
> Reference Parameters directly as first class SOAP header 
> blocks (with no 
> change). This implies that SOAP 1.1/1.2 attributes such as 
> soap11:mustUnderstand, soap12:mustUnderstand, soap11:actor, 
> soap12:role, 
> soap12:relay must be specified in the value of Reference 
> Parameters in 
> an EPR instance. I.e. the abstract property [Reference 
> Parameter] value 
> can be SOAP 1.1/1.2 specific. This means that when specifying a value 
> for a reference parameter the minter of the EPR has to know 
> not only the 
> fact that the EPR is being used with SOAP but the version of 
> SOAP as well.
> 
> Going through the spec the only place that it talks about 
> abstractness 
> is section 2.1 [1].
> 
> But this abstractness may only have to do with the fact that it is 
> independent of serialization. I (and I suspect Glen as well) had 
> interpreted WS-Addressing abstract properties to mean that 
> all instance 
> of an EPR could be bound to either SOAP 1.1 (using SOAP 1.1 
> binding) or 
> SOAP 1.2 (using SOAP 1.2 binding) or some other protocol. If this 
> interpretation is correct then the SOAP binding violates such an 
> abstractness since it would inject soap version specific 
> MU/role/actor/relay attributes in the value of the Ref 
> Params. If this 
> interpretation is incorrect, i.e., the abstractness has to do 
> with the 
> fact that it is independent of serialization then it seems like there 
> isn't an issue here.
> 
> Comments?
> 
> -Anish
> --
> 
> [1] 
> http://dev.w3.org/cvsweb/~checkout~/2004/ws/addressing/ws-addr
> -core.html
> 
>

Received on Monday, 7 February 2005 12:34:55 UTC