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

Re: HTTP binding options

From: Sanjiva Weerawarana <sanjiva@watson.ibm.com>
Date: Mon, 10 Nov 2003 19:58:46 +0600
Message-ID: <07e101c3a792$c3f59020$ebc16720@lankabook2>
To: "FABLET Youenn" <youenn.fablet@crf.canon.fr>, "Philippe Le Hegaret" <plh@w3.org>
Cc: "Web Services Description" <www-ws-desc@w3.org>

"FABLET Youenn" <youenn.fablet@crf.canon.fr> writes:
> I do not remember that there was a "pretty strong sentiment against 
> doing 5", i.e. URL replacement.

I was referring to the lunch table discussion .. at which, IIRC, you 
were not there. Sorry, should've been clearer.

> Maybe I do not recall the entire discussion. Anyway, I would also favor 
> option 5, which seems to be equivalent to today's http binding 
> functionnality.

It is, but today's only works like that for messages with multiple parts
where the parts are all simple. We'd need to effectively define a
specialization of the RPC style to make it work now .. thereby 
bleeding @style to bindings, which would be unfortunate IMO. That's
what lead to most of the WSDL 1.1 interop problems I believe.

> Personly, I would even go beyond and ask to generalize the access 
> mechanism (used by the url replacement) to work not only with the http 
> binding but also with the soap binding, for instance to directly set 
> property values with abstract data.

That makes the binding language into a pretty powerful language for
how an HTTP request is formed. I think that's overkill for what we need.

Sanjiva.
Received on Monday, 10 November 2003 08:56:57 GMT

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