RE: On WSDL "operation"

+1 to Sanjiva and Jim.

One might be tempted to remain neural on this point -- it could be
messages or it could be objects -- but we must be very clear: it is
messages, and you may build whatever programming model you choose. This
is a very germane point when the WG is tempted to eliminate or limit
constructs in WSDL that are not useful for object serializations.

--Jeff

> -----Original Message-----
> From: www-ws-desc-request@w3.org [mailto:www-ws-desc-request@w3.org]
On
> Behalf Of Jim Webber
> Sent: Monday, September 29, 2003 9:05 AM
> To: 'Sanjiva Weerawarana'; 'Mark Baker'
> Cc: 'WS Description List'
> Subject: RE: On WSDL "operation"
> 
> 
> Sanjiva:
> 
> > It defines documents which some may choose to interpret as object
> > serializations. What's the problem with that?
> 
> I don't have a problem with that, and indeed that is the canonical
case.
> There are some documents defined somewhere, and in my programming
> environment I have a view onto them in a form that suits me. It might
be
> objects, it might be something else.
> 
> > "Abused" is a strong word for describing an interpretation that
> > most implementations have of the XML documents that are described
> > in WSDL: eventually these become data structures in a programming
> > language and a Java object is a lot more Java programmer friendly
> > than a DOM tree. That does not make WSDL an object IDL.
> 
> I disagree. What we have seen in Web Services is a tendancy to expose
> objects via WSDL which is abuse, plain and simple. The fact is that
most
> developers just aren't thinking in document centric terms, and
continue to
> equate methods on objects to WSDL operations. And there are examples
out
> there of exactly this abuse going on.
> 
> I hope that WSA will clarify this, and we can move on. If WSDL grasped
the
> bull by its horns and used the right terminology and embraced the
"simple
> message exchange model" it would be a wonderful help in educating
> developers
> in how to build service- rather than object-oriented systems.
> 
> Jim
>

Received on Wednesday, 1 October 2003 23:41:15 UTC