RE: Mandator wsa:Action (was Re: WS-Addr issues)

 

> -----Original Message-----
> From: public-ws-addressing-request@w3.org 
> [mailto:public-ws-addressing-request@w3.org] On Behalf Of Marc Hadley
> Sent: 05 November 2004 15:46
> To: David Orchard
> Cc: Francisco Curbera; public-ws-addressing@w3.org; Mark Little
> Subject: Re: Mandator wsa:Action (was Re: WS-Addr issues)
> 
> 
> On Nov 4, 2004, at 4:52 PM, David Orchard wrote:
> >
> > So you are arguing that action may or may not be RPC.
> 
> I'm arguing that the presence of action is orthogonal to the 
> programming model. One can build an RPC-like mechanism based on 
> dispatching on action or on the message payload. Similarly one can 
> build a message oriented mechanism on either. The presence or lack of 
> action doesn't mandate a particular programming model.
> 
> > Without going
> > further on that (which I could but we've got an overload of messages
> > already), my point was that people ended up always looking into the
> > message to determine the "action" under the optional soap 1.1 action
> > header.  A mandatory WSA:Action breaks that cycle and an optional 
> > Action
> > perpetuates it.
> >
> I'm OK with a particular service requiring the presence of an action. 
> I'm not OK with requiring every message to carry one even when the 
> service they are destined for doesn't use it. This is where 
> we ended up 
> in the XMLP WG and I think its a good compromise position.

If a service doesn't require wsa:Action, then perhaps it shouldn't be
usign WS-Addressing?

Gudge