- From: Diane Mularz <mularz@mitre.org>
- Date: Fri, 27 Apr 2001 14:41:55 -0400
- To: Amit Bhatiani <amit@invertica.com>
- CC: www-ws@w3.org
- Message-ID: <3AE9BD73.AE887E9A@mitre.org>
Amit Bhatiani wrote: > > Hugo Haas said: > > >I personnally think that describing semantics of the service at the > >same time as its operation makes sense. > > 1. http://www.w3.org/TR/2001/NOTE-wsdl-20010315#_documentation > 2. http://www.w3.org/2001/03/WSWS-popa/paper51 > > I agree with you on describing semantics (and in fact transactional > semantics if needed) as first class objects in WSDL along side descriptions. > I just don't think that "meaning of methods" and semantics of web services > are interchangeable terms. The web service might have semantics that are > defined at multi-method level only and not at each method. In fact, some > methods might make no sense to call unless others are called before/after. > Method signatures should also be included along with invariants, but is that > the same as WS Semantics? Perhaps one needs to distinguish between 'service' and 'method'. A service in my mind performs some business necessary processing (e.g., submit order). The service request may actually result in a series of method calls on one or more underlying components to provide the service. However, all of this should be transparent to the invoker of the service. So the service description would be defined at a business level and there is perhaps an implementation description that defines how it is satisfied by any implementer of the service. Or perhaps there are hierarchies of services beginning at low level primitives and leading to higher order, business-oriented services. An implementer of any given service would then specify the service in terms of other services lower in the hierarchy. > > ______________________________ > 212-571-4103 x 6590 (w) > 201-759-4158 (m) > 212-571-3588 (f) > > Invertica Inc. > www.invertica.com
Received on Friday, 27 April 2001 14:43:15 UTC