RE: proposal for eliminating message

I'd like to know two things:

1. What does this look like at the component model level? This would
help me to evaluate the expressive power vs simplicity/complexity.

2. Why do we continue to re-invent constructs that already exist in XML
Schema? 

I think references to named model groups would suffice here. References
to types or global element decls would also be fine. All three allow one
to address ordering, choice and cardinality with existing mechanisms.

I propose we decide to allow some combination of type/element/group
attributes on input/output/fault. And we specify that other type systems
can put qualified attributes on those elements to do their thing. Such
an approach would make the following changes to the component model:

1.	replace the {message} property of the Message Reference
component[1] with the {content reference} property from the Part
component[2]. 

2.	Remove the Message component.

3.	Remove the Part component. 

It would also make sense to change the name of the Message reference
component.

Gudge

[1]
http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/wsdl12.xml#Part_
details
[2]
http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/wsdl12.xml#Messa
geReference_details

> -----Original Message-----
> From: Sanjiva Weerawarana [mailto:sanjiva@watson.ibm.com] 
> Sent: 18 January 2003 00:30
> To: www-ws-desc@w3.org
> Subject: proposal for eliminating message
> 
> 
> Attached is an attempt at a compromise proposal for removing 
> the <message> construct.
> 
> Sanjiva.
> 
> 

Received on Monday, 20 January 2003 06:47:36 UTC