RE: Thoughts on TAG issue EndpointsRef47

I agree with your position. But I think the spec should have some
clarifying words in this area, particularly in the case where there are
SOAP intermediaries (issue that has already been raised before).

In particular, in the presence of SOAP intermediaries:

- what is "wsa:To" referring to: the ultimate receiver or one of the
intermediaries? (the former, I believe, but it should be spelled out)

- how does the "wsa:To" URI relate to the individual physical addresses
involved, i.e. the addresses of each intermediaries and the address of
the ultimate receiver? (if, as I think, there is no necessary relation,
we should at least clarify whether there are ways to carry those
physical addresses in the SOAP envelope using WS-A if so desired)

- how does the URI scheme of "wsa:To" relate to the transport(s)
involved? (no relation in the general case, I would say)


> -----Original Message-----
> From: 
> [] On Behalf Of 
> Savas Parastatidis
> Sent: Tuesday, February 08, 2005 3:45 PM
> To:
> Cc:
> Subject: RE: Thoughts on TAG issue EndpointsRef47
> > 
> > Why isn't a http URL logical? It is a logical address that 
> happens to
> be
> > resolved by an application-specific combination of DNS 
> lookup and Web 
> > Server at message transfer time. Savas' prior example of using a URN
> to
> > address a service hosted on a P2P network is no different really.
> > 
> To further support Jim's argument above, let me give another 
> example. Do you think that WS-Addressing should specify that 
> when the SOAP message bellow is sent by a sender the sender 
> MUST use HTTP for the transfer of that message?
> <soap:Envelope>
>   <soap:Header>
>     <wsa:To></wsa:To>
>     <wsa:Action>urn:ProcessMessage</wsa:Action> <!-- :-) -->
>   </soap:Header>
>   <soap:Body>
>     <!-- bla bla -->
>   </soap:Body>
> </soap:Envelope>
> What if the original sender decided to use a pigeon to 
> transfer this SOAP message to an intermediary? Does it really 
> have to use HTTP? Should WS-Addressing define the semantics 
> on how a SOAP message is transported?
> Regards,
> .savas.

Received on Wednesday, 9 February 2005 03:24:27 UTC