W3C home > Mailing lists > Public > www-ws-desc@w3.org > July 2003

RE: eliminating <message>: a few additional thoughts

From: Savas Parastatidis <Savas.Parastatidis@newcastle.ac.uk>
Date: Mon, 21 Jul 2003 20:13:22 +0100
Message-ID: <BC28A9E979C56C44BCBC2DED313A447001EC32A0@bond.ncl.ac.uk>
To: <www-ws-desc@w3.org>

> 
> True. However, that's make the simple, what I believe to the 80-20
> case harder. (See near the top of my original proposal and you'll
> see what it looks like.)
> 

I believe that in most cases the SOAP body will have more than one
element. Example:

<xs:complexType name="FooBody">
    <xs:sequence>
        <xs:element name="Bar1" type="xs:string"/>
        <xs:element name="Bar2" type="xs:positiveInteger"/>
    </xs:sequence>
</xs:complexType>

<xs:element name="FooElement" type="tns:FooBody"/>

<interface name="MyInterface">
   <operation body="Foo">
      <input body="tns:FooBody"/>
   </operation>
</interface>

Why do you see this example as the 20% of the cases? Using the element
approach that would be...

<interface name="MyInterface">
   <operation element="Foo">
      <input element="tns:FooElement"/>
   </operation>
</interface>

Effectively this means that you have a wrapper element, called
"FooElement" inside body.

I am not particular bothered which one is adopted at the end but I do
have a small preference to the former :-)

> 
> I'm still debating about how to do @body, but the optionality
> is ok with me. I seem to recall that SOAP requires a non-empty
> body, but I have seen specs which don't seem to follow that.
> 

Actually the SOAP 1.2 spec allows non-empty bodies... (don't remember
whether 1.1 allowed this)...

Section 5.3 of the spec:

<quote>
The Body Element information item has:
...
- Zero or most namespace qualified element information items in its
[children] property.
...
</quote>

I hope this helps.

.savas.
Received on Monday, 21 July 2003 15:13:45 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:25 GMT