- From: Glen Daniels <gdaniels@sonicsoftware.com>
- Date: Thu, 8 Jan 2004 13:00:54 -0500
- To: <paul.downey@bt.com>, <www-ws-desc@w3.org>
Hi Paul! > As I now understand it, in WSDL 2.0 the schema employed to > describe the data being exchanged will also describe the > actual serialisation of the data 'on the wire'. This is not in fact the case. The schema describes the infoset which will be "rehydrated" at the receiving end, no matter the serialization. Depending on where we end up going with things like soap:header, that might not even exactly be true either (though I hope it remains so). > Doesn't this remove the ability for the same infoset to be > exchanged using different serialisations: XML text, MTOM, or > whatever and necessitate a new schema /language/ just to > support a different message encoding ? I don't believe so, no. > It seems to me that there was real benefit in encodingStyle: > multiple bindings could easily present the same data but > exchanged over different transports *and* serialisations. The benefit to encodingStyle was that you could rely on the schema looking a particular way (elements, no attributes, etc) and that you could know that any given element's content might be replaced with an href="" attribute. --Glen
Received on Thursday, 8 January 2004 13:03:50 UTC