Re: http binding

Little work has been done so far on the SOAP binding. Certainly, 
your suggestions fit with the way I see the SOAP binding moving 
forward, i.e. a transport agnostic binding, the MEP being 
selected via the WebMethod feature, this feature appearing either 
at the binding or (binding-)operation level.

As far as fantasy is concerned (a unified HTTP-SOAP binding), 
this may set the bar too high.

Jean-Jacques.

Sergey Beryozkin wrote:

[snip]

> My understanding was that WSDL SOAP binding is a specific implementation of
> an abstract SOAP HTTP binding as defined by SOAP Adjuncts (sorry if the
> terminology is not right). Web Method Feature can tell whether SOAP HTTP
> binding allows for SOAP Request-Response MEP or SOAP Response MEP
> interaction style.
> What I'm not sure about is that whether Web Method Feature can be applied in
> a fine-grained fashion, on a per-operation basis or not. I thought it would
> be applied to the whole binding instance. Another issue is that GET does not
> allow for SOAP headers, so GET can't be used while SOAP Request-Response MEP
> is active.
> But if it were possible to use GET with headers [2] for idempotent
> operations, then, may be, WSDL SOAP binding could provide for an abstract
> SOAP HTTP binding + SOAP Request-Response MEP, with SOAP Response MET being
> redundant, as  SOAP Request-Response MEP could use GET and POST
> interchangeably. Probably, this sounds terribly wrong :-), but if it were
> possible at least in theory, then  WSDL SOAP binding would look quite
> similar to WSDL HTTP binding (at least with doc-lit style in place), and if
> it were the case, then why don't merge the two ? Fantasy :-)
> Just my 2c
> Thanks
> Sergey Beryozlom