W3C home > Mailing lists > Public > www-ws-desc@w3.org > December 2002

Re: Outboud ops state of debate

From: Jean-Jacques Moreau <moreau@crf.canon.fr>
Date: Wed, 11 Dec 2002 17:22:55 +0100
Message-ID: <3DF7665F.9030209@crf.canon.fr>
To: Jonathan Marsh <jmarsh@microsoft.com>
CC: www-ws-desc@w3.org, FABLET Youenn <fablet@crf.canon.fr>

As indicated on the call, I have reservations[1] with publishing 
the TCP binding as is. I don't think this is quite the way you 
would want to write a SOAP TCP binding from a  SOAP perspective. 
I am fine with publishing an IRC/XML to TCP binding, along the 
lines of what Jeff is  proposing, but I have an issue with 
providing a  SOAP binding in a non SOAP 1.2 conformant way. I 
think this would  send the wrong message and be an unfortunate 
precedent.

To explain a bit more, there is not just one but two bindings at
stake here. Jeff has provided a *WSDL-binding* for IRC over SOAP
over TCP. The binding describes how you would map IRC commands to
SOAP messages and then to TCP. I don't have any problem with the
IRC to SOAP part. However, I have reservations with the second
part, SOAP to TPC, and this is where the second binding comes
into play, a SOAP to TCP *SOAP-binding*.

SOAP 1.2 currently provides a SOAP to HTTP SOAP-binding only.
This binding specifies how to map SOAP envelopes to HTTP messages
and then how these messages are exchanged and processed. In
addition, SOAP 1.2 provides a framework for describing other
SOAP-bindings. For example, one might wish to provide a SOAP to
TCP SOAP-binding. Such a binding would indicate how to transport
SOAP envelopes directly over TCP. An implementation of this
binding would deal connection establishment, framing, etc. From a
programmer's perspective, however, there would likely be little
difference between this SOAP-binding and the existing HTTP
SOAP-binding: both would transport SOAP envelopes.

If such a TCP SOAP-binding was available, we would just be able
to use the standard SOAP WSDL-binding, as follows:

<soap:binding uri="http://example.org/2002/12/soap/bindings/tcp"/>

instead of Jeff's current proposal:

<stp:binding framing="http://example.org/2002/10/soap/tcp/dime"/>

So, to summarize, I am relunctant to publishing the current TCP 
binding as is, as I think it would send the wrong signal. I think 
it's an excellent first step, but that we need to fix the 
SOAP-binding issues before publication. (I would also like to 
note that Youenn has suggested to me that there are potentially 
some MEP issues here as well.)

I would suggest that we either rework the proposal to not use
SOAP or that we rely on a (yet-to-be-written) TCP SOAP-binding.

Does this make sense?

Jean-Jacques.

[1] http://lists.w3.org/Archives/Public/www-ws-desc/2002Nov/0051.html

Jonathan Marsh wrote:
> 2) Publish the TCP binding as a note motivating the need to retain
> output operations.
Received on Wednesday, 11 December 2002 11:23:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:22 GMT