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

Re: RPC Style Issues (3)

From: Jacek Kopecky <jacek.kopecky@systinet.com>
Date: Thu, 30 Oct 2003 16:15:09 +0100
To: Umit Yalcinalp <umit.yalcinalp@oracle.com>
Cc: Sanjiva Weerawarana <sanjiva@watson.ibm.com>, Tom Jordahl <tomj@macromedia.com>, 'WS Description List' <www-ws-desc@w3.org>
Message-Id: <1067526909.3354.21.camel@localhost>

Umit, others,

I can see how the rpc:result attribute would work and I don't think it
would be hard to spec or implement. Therefore, I think I can support it.

Anyhow, this whole issue only comes up in a few scenarios which IMHO
fall below the 80/20 line:

     1. round-trip interop - if I can't distinguish the return value and
        the first out parameter (if it's the first parameter), a
        language interface generated from a WSDL generated from a
        language interface may change.
     2. support for interfaces containing two operations named the same,
        one with a return value and one with an out param of the same
        type - WSDL interface won't be able to keep that

Otherwise, not having the distinction doesn't harm interop.

BTW, the reason behind SOAP 1.2 RPC's indirect identification of the
return value is that a SOAP Data Model struct doesn't have any notion of
order and therefore there is no such thing as *the first* out parameter.
We didn't mean to introduce this distinction, it was a byproduct, IIRC.

When replying please remember I said in the beginning of this message 
I can support the attribute. 8-)

                   Jacek Kopecky

                   Systinet Corporation
                   http://www.systinet.com/
Received on Thursday, 30 October 2003 10:15:14 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:27 GMT