Re: HTTP binding options

On Mon, Nov 10, 2003 at 03:40:49PM -0500, Champion, Mike wrote:
> Uh, no, it just means that we need to ensure that WSDL can describe services
> that use the REST constraints, which is a small subset of the services that
> WSDL can describe.

Didn't I say that?  I meant to.

>   A constructive proof of that is that WSDL can (I think)
> describe HTTP itself.  

I agree that it's a proof, but FWIW, WSDL 2.0 cannot describe HTTP for
the reason I mentioned; WSDL 2.0 has self-description issues regarding
its failure to permit services to declare which symbol in the message
any success or failure in the processing of that message corresponds to
(as yet another way of explaining the ambiguity issue I raised).

But that isn't specific to HTTP.  WSDL 2.0 can't currently unambiguously
describe *any* service.  The attempt to have its cake and eat it too
(i.e. trying to graft the loose coupling aspects of "document exchange"
onto RPC) has left it suitable for neither.

Mark.
-- 
Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca

Received on Monday, 10 November 2003 16:09:55 UTC