- From: Savas Parastatidis <Savas.Parastatidis@newcastle.ac.uk>
- Date: Wed, 10 Sep 2003 18:01:00 +0100
- To: "Steve Graham" <sggraham@us.ibm.com>
- Cc: <public-ws-desc-state@w3.org>
Steve, > > Re: fundamental concepts only. > This is a very slippery argument. Do *all* web services need the > sophisticated MEPs suggested by the MEPs work? No. Do *all* Web services > need properties and Features? No. There are many things in WSDL that are > not needed by *all* Web services, but are added for reasons of common > interoperability. Our claim is that attributes fit into this category. I > understand that you and Savas do not concur. > Our argument is that attributes introduce another way to describe a message exchange, and are therefore superfluous. Certainly properties and features may not be required by all web services, but they do not simply provide an alternative means of doing something that is already achievable. In cases where a declarative approach to defining requirement/support for WSA protocols, properties/features is the WSDL-provided mechanism. The same with MEPs. They address a specific requirement. In the updated proposal, attributes describe a message exchange which is already possible with existing mechanisms in WSDL. Furthermore, as we suggested earlier, since attributes is not a concept defined by WSA, it's going to further confuse the community about what a service is since we are mixing implementation and architectural concepts. Is it an object with methods and attributes or an entity that just sends/receives messages? We believe that the current mechanisms in WSDL suffice for describing Web Services. Finally, disagreement is good. If we all agreed on everything, there wouldn't be much point in discussions. No? ;-) Regards, Savas and Jim
Received on Wednesday, 10 September 2003 13:02:03 UTC