RE: Proposal for dealing with root, top-level multi-refs and encodingStyle

Henrik Frystyk Nielsen writes:

>> I think the only need for introducing the notion 
>> of a root in the data model is to describe 
>> that whenever you have cross-references between
>> serialization trees, you have a single graph.

I guess I don't see why the notion of root is the right way to achieve 
that.  If that's what we're trying to say, why not include the following 
clarification:

"All element information items representing serialized nodes and envelopes 
must be within the scope of  an encodingStyle="...soap encoding URI..". 
The number and placement of such encodingStyle attributes is not otherwise 
significant;  ref attribute information items MAY refer to ID's in the 
scope of a same or different encodingStyle attribute, which MAY but need 
not be within the same SOAP body or header block.  In all such cases, the 
result is a single edge (and therefore a single graph.)"

The wording might be cleaned up, but I think that's clearer than 
introducing a notion of root. 

------------------------------------------------------------------
Noah Mendelsohn                              Voice: 1-617-693-4036
IBM Corporation                                Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------

Received on Tuesday, 7 May 2002 14:52:55 UTC