Re: Section 5 vs Schema

I think the real question is about the status of the graph model.  Schema 
models data as elements and attributes.  Chapter 5 of SOAP 1.1 models data 
as nodes carrying typed values, connected by named edges.  Chapter 5 also 
describes serialization and deserialization of graphs using XML.   Chapter 
5 is motivated, I think, by a perception that many programming structures 
(DAGs, Trees, Lists, etc.)  are better modeled as graphs.  Furthermore, 
many languages distinguish "by reference" from "by value", leading to 
graphs in which multiple edges terminate in a single node. 

The relationship between a given graph and an XML schemas describing 
serializations of that graph is one-to-many.  There are many 
(interestingly) different legal schemas that would all validate some legal 
serialization or another of what you would consider a single graph schema. 
 WSDL has a particular spin on this problem, though it's not one I 
paritcularly like. 

Anyway, I think that investigating Gudge's proposal boils down to 
re-evaluating why we need a graph model, and if we do need it, what 
characteristics the schemas describing those graphs should have.

------------------------------------------------------------------------
Noah Mendelsohn                                    Voice: 1-617-693-4036
Lotus Development Corp.                            Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------------







Francis Norton <francis@redrice.com>
Sent by: xml-dist-app-request@w3.org
12/04/01 01:02 PM
Please respond to francis

 
        To:     Martin Gudgin <marting@develop.com>
        cc:     XML Protocol Discussion <xml-dist-app@w3.org>
        Subject:        Re: Section 5 vs Schema


Hi,

Martin Gudgin wrote:
<lots of sensible stuff>


It would be great to have some explanation to this effect in some part
of the final document - coming from a very schema-oriented background I
really didn't understand the place or agenda of the old section 5, and
the way it was being put forward without what I regarded as a strong
rationale made me feel kind of prickly, like it was out to reinvent XML
Schema.

By contrast Martin's explanation makes me feel downright warm and fuzzy...

Francis.

Received on Thursday, 6 December 2001 12:58:23 UTC