If you are overriding the element type in the schema instance (as in your example), then it's appropriate to use the xsi:type attribute in a doc/literal SOAP message. But certain SOAP engines, such as .NET, don't take kindly to messages containing the xsi:type attribute.
Anne ----- Original Message ----- From: "Jeff Greif" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Monday, August 04, 2003 11:50 AM Subject: Re: WRAPPED services without wsdl > Could you clarify why you don't think attributes such as xsi:type should be > valid in doc/lit. > > Suppose your document is described by > > <element foo type='bar'/> > > <complexType name="bar" abstract="true"> > ... > </complexType> > > <complexType name="bar1" > <extension base="bar"> > ... > </extension> > </complexType> > > ... other types derived from bar ... > > Then you might have a document > <foo xsi:type="bar1">...</foo> > > where the xsi:type told which derived type was being used. > > Jeff > ----- Original Message ----- > From: "Cory Wilkerson" <[EMAIL PROTECTED]> > To: <[EMAIL PROTECTED]> > Sent: Monday, August 04, 2003 8:33 AM > Subject: RE: WRAPPED services without wsdl > > > I guess my question is -- is xsi still valid in a doc/lit operation? It > doesn't seem like it would be. >
