My issue with WSDL is in the mixing of technology with abstract specification and with the clumsy way it expresses things partially because it doesn't support the concept of an object. There is an OMG specification to translate between IDL and WSDL, so I don't think there is something that can't be expressed in either. An example of the clumsiness of WSDL is that a "simple" 200 line IDL file turns into more than 3000 lines of WSDL. The 200 line IDL file is human readable and understandable. The equivalent WSDL file is almost impenetrable.
Dave Bill de hOra wrote: > David Forslund wrote: > > >> WSDL is a very >> poor way of expressing architectural abstractions. That is what UML is >> for. IDL is a little better >> for architectural abstractions, primarily because it is somewhat higher >> level than WSDL. The equivalent >> expression in WSDL of an IDL file will typically be more than 10 times >> longer. Some of this is because >> WSDL has no way of expressing the semantics of objects (which frequently >> translate nicely into services). >> > > I agree WSDL has issues with description, but saying IDL has a semantics > is odd to me; all it does is define signatures using a type system. Did > you mean OCL? Or do you have an example of something that WSDL can't say > that IDL can? > > cheers > Bill > >
