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

Re: On WSDL "operation"

From: Mark Baker <distobj@acm.org>
Date: Mon, 29 Sep 2003 12:26:58 -0400
To: Jim Webber <jim.webber@arjuna.com>
Cc: WS Description List <www-ws-desc@w3.org>
Message-ID: <20030929122658.E9630@www.markbaker.ca>

BTW, Jim, I agree with your response to Sanjiva that this is more than
just a naming issue

On Mon, Sep 29, 2003 at 09:30:05AM -0400, Jim Webber wrote:
> Mark:
> 
> > I think I understand what you're asking for, but it's something very
> > different than what WSDL currently is and does.  With your proposal,
> > what WSDL is currently used for would require using a technology such
> > as IDL.
> 
> I believe that WSDL has been abused as an object IDL, and I have abused it
> in this way myself, and toolkit vendors help propagate that view. However,
> objects and Web Services are chalk and cheese, and WSDL should evolve to
> take into account the fact that we as a community are discovering how to
> properly use Web Services (i.e. that the semantics are simply that they
> exchange messages and do not share type information).

Fair enough.

> If I have (inadvertently) tabled a proposal, that it is that we should
> simplify the basic view of Web Services to be just entities exchanging
> messages (which is WSA, no?), and that WSDL should support that and provide
> extensibility mechanisms for all the other stuff. Other higher-level Web
> Services protocols can be layered on top of this simple underlay.
> 
> So in short:
> 
> 1. Web Services exchange messages.
> 2. WSDL describes those messages (and perhaps how they might be exchanged).
> This includes both abstract and concrete forms of those messages.
> 3. All the other stuff is out of scope (and indeed only makes sense when
> there is an application to resolve what it means).
> 
> I believe WSDL can do this, it's just that with nouns like "operation" we
> implicitly suggest to developers that WSDL is an IDL, when it isn't, it's a
> CDL.

It seems that if the group has decided to keep the name "operation",
that the semantics should be the same too, no?  This would mean that the
answer to my question[1] is "the former", that a successful response
means that the foo operation was invoked.  I find it very difficult to
reconcile your position with this answer, which suggests that the answer
to my question may be "either" and therefore the interface/contract
ambiguous.

> PS - Does this mean we agree? :-)

Hah, funny! 8-)

FWIW, I have a lot of sympathy with this view as we previously
discussed.  But our approaches are quite different.  I believe that
Web services should use a loosely coupled document exchange
architecture, where the "operations" are specific to document
exchange (aka "uniform").

 [1] http://lists.w3.org/Archives/Public/www-ws-desc/2003Sep/0194.html

Mark.
-- 
Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca
Received on Monday, 29 September 2003 12:22:37 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:26 GMT