- From: David Hull <dmh@tibco.com>
- Date: Wed, 08 Nov 2006 12:02:08 -0500
- To: paul.downey@bt.com
- Cc: curbera@us.ibm.com, public-ws-addressing@w3.org, public-ws-addressing-request@w3.org
- Message-id: <45520D90.9010507@tibco.com>
Paul, That's certainly an interesting way to phrase the question, and I wouldn't mind seeing a transcript of the SOA thing, but I'm just after a good old-fashioned contract. For example, we know what a conforming endpoint will do when it receives a request message with WSA headers, both in terms of what the headers in the response will look like, and in terms of what it will do when the response endpoints contain WSA anon and/or WSA none. In particular, we know that an endpoint that accepts WSA anon as a response endpoint has to support the SOAP request-response MEP. We know that there is no constraint on behavior in the face of a non-unique message ID. That may seem like a lack of a contract, but (at the risk of sounding like Donald Rumsfeld), we know that we don't know what the endpoint will do in that case (unless we do know, through other means :-). All I know about a back-channel is that it's somehow related to WSA anon, that no one defines it right now, and that a binding can define it however it wants. IMHO, that's not a contract I can write code to. paul.downey@bt.com wrote: > >> In order to have a meaningful >> definition of X functionality, we need a clear answer to "If something >> says it provides X, what can I count on?". >> > > I think you are asking for an "existence" or "nonconstructive proof" > i.e. if I didn't have a back channel, how would I know? > > I heard Sean McGrath ask an audience the same question about "SOA" > with hilarious results .. > > > > >
Received on Wednesday, 8 November 2006 17:02:36 UTC