>Are you aware of any other vendors that use the "wrapped" style?  Don't you
>think that using a part name (and the fact that there is only one) as a
>means of discriminating between "document" and "wrapped" is fairly delicate?
>I appreciate that this might be pedantry, but what if I wanted "document",
>not "wrapped" and named the single part "parameters"?

> I believe that the IONA folks are supporting this as well. I agree
> with you that keying off the part name is a pretty flaky approach.

What would you think we should key from?  This is fairly reliable right now as this is 
how Microsoft (and others?) are creating the WSDL.

It is likely that in the future we will provide a WSDL2Java switch to turn off this 
heuristic
and just do straight-up doc/lit stuff.

--
Tom Jordahl
Macromedia




-----Original Message-----
From: Simon Fell [mailto:[EMAIL PROTECTED]]
Sent: Thursday, April 04, 2002 9:03 PM
To: [EMAIL PROTECTED]
Subject: Re: Wrapped vs. Document


On Thu, 4 Apr 2002 13:55:45 +0100, in soap you wrote:

>Are you aware of any other vendors that use the "wrapped" style?  Don't you
>think that using a part name (and the fact that there is only one) as a
>means of discriminating between "document" and "wrapped" is fairly delicate?
>I appreciate that this might be pedantry, but what if I wanted "document",
>not "wrapped" and named the single part "parameters"?

I believe that the IONA folks are supporting this as well. I agree
with you that keying off the part name is a pretty flaky approach.

>This leads me to a slightly deeper question (which is connected, however
>tenuous the link may seem at first), which has likely been debated before,
>so I apologise if this is ground already covered:
>
>Microsoft obviously use document style for the majority (all?) of there
>services.  This means that the data in the SOAP packets which they send is
>untyped on the wire, putting the emphasis on the framework to
>marshall/unmarshall based on the contract (WSDL), whether statically or
>dynamically.  I guess the question is, why bother to encode (using SOAP
>section 5 rules)?  Surely a combination of document style and XML schema
>(which is also a means of describing serialisation/deserialisation rules for
>XML) is sufficient for all cases?

section 5 provides a standard serialization for graphs, which XML
schema does not. At this point, its the only useful thing in section
5.

Cheers
Simon
www.pocketsoap.com

Reply via email to