> That's OK. > This type of thing would not make 1.0 as we are in the process of > closing down.
Yep, noticed. Very exciting! > Note that this work is non-trivial. It involves delaying the > typing of data till after you receive it based on the contents of > the schema in the data. Ok. Questions: - is that a required part of the SOAP spec or just a convenience? I haven't noticed it (not that I completely understand the spec). - if it's part of the spec, why wouldn't you at least add support for recognizing the xsd:schema type now for compliance, even if it is implemented as, say, 'any'? The caller can at least get the schema document back then. - what exactly is it that you think it should do (or, asked a different way, what are the integration points with Axis)? + should the typing be done dynamically such that the stubs instantiate the proper object type but don't declare it? + should the typing be done by WSDL2Java such that it's embedded in the stubs + should it be a 'casting'/'validating' layer on top of the WSDL2Java generated stubs? + other? Per
