- From: Don Box <dbox@microsoft.com>
- Date: Wed, 6 Feb 2002 17:10:01 -0800
- To: <noah_mendelsohn@us.ibm.com>, <soap@zaks.demon.co.uk>
- Cc: <xml-dist-app@w3.org>
> -----Original Message----- > From: noah_mendelsohn@us.ibm.com [mailto:noah_mendelsohn@us.ibm.com] > Sent: Wednesday, February 06, 2002 3:31 PM > To: soap@zaks.demon.co.uk > Cc: xml-dist-app@w3.org > Subject: Re: Issue with soap-rpc:result > > FWIW, RPC in SOAP 1.1 technically depends on the data model that's > provided by the encoding. It models arguments as a struct [1], which is a > concept that only exists in the encoding. This approach was designed, in > part, to leave the door open to other encodings that would also model > structs, but as with so much in SOAP 1.1 the spec is quite broad in its > language, and I think you can interpret it as either requiring the chapter > 5 encoding, or just requiring some encoding that models "structs". > > We surely need a less ambiguous presentation in SOAP 1.2. Whether to > allow completely unencoded (E.g. modeled by XML schema or with no schema, > rather than being a graph with structs, etc.) I am less sure. Since SOAP/1.2 explicitly relies on XML Schema, I would advocate making faults based solely on XML Schema (e.g., literal). This is especially important since the fault itself may be indicating that an unknown encoding was used! DB
Received on Wednesday, 6 February 2002 20:10:49 UTC