- From: <paul.downey@bt.com>
- Date: Thu, 1 Jul 2004 13:04:11 +0100
- To: <hugo@w3.org>, <dbooth@w3.org>
- Cc: <www-ws-desc@w3.org>
Hugo wrote: > Thank you for this excellent analysis. I think that you framed > the problem very well. yes, thanks David, excellent! > Therefore, thinking out loud, I think that we could say in Part 1 > that, if the GED of operations is not unique, extra information must > be used in order to unambiguously identify the operation relating to a > message received by a service, by using a feature at the interface > operation level. We could even define such a feature, say the > Operation Name Specification Feature, which would be an abstract > feature: we would not define how it is implemented (in SOAP, HTTP, or > any other way), just noting that its use makes the receiver of the > message aware of which operation the message is related to. So the default for an absent feature (or do i mean property?) would be routing via the GED, but this may be overridden to point to another mechanism? Either way the publisher of a service would have to explicitly declare in WSDL the dispatching mechanism being employed. That sounds reasonable to me. Paul
Received on Thursday, 1 July 2004 08:04:14 UTC