Amy Lewis wrote:
I am assuming that client code can get at the WSDL that defines the
service and the interface defined by those two URIs. Then the client
code can, at run-time, determine what transport to use and then use
it. It is useful to determine the transport only at runtime, since
that allows the service provider to change the transport without
breaking references that clients have to their service.
On Wed, 16 Jul 2003 17:21:06 -0400
Michael Rowley <firstname.lastname@example.org> wrote:
Great example. Now assume that Ma & Pa want to advertize a URI for
subscribing to their service _without_ implying the binding that will
be used in the URI. How do they do that? You obviously can't send
out any of the soap:address locations, since they imply a transport.
You can't send the URI of the "ma-pa-market-service" element, since
that is for more than one interface. Seems you would need to send out
two URIs -- the one for "ma-pa-market-service" and the one for the
"ma-pa-market-admin" interface. Two URIs are necessary in order to
refer to the most useful concept that is provided by WSDL: a reference
to a service independent of the mechanism used to communicate with it.
Why is that the most useful concept? What can I do with these two URIs?
If I want to subscribe ... I need a URL. No?