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

RE: On WSDL "operation"

From: Savas Parastatidis <Savas.Parastatidis@newcastle.ac.uk>
Date: Mon, 29 Sep 2003 17:19:52 +0100
Message-ID: <BC28A9E979C56C44BCBC2DED313A4470021C2BC9@bond.ncl.ac.uk>
To: "Sanjiva Weerawarana" <sanjiva@watson.ibm.com>
Cc: "WS Description List" <www-ws-desc@w3.org>

Sanjiva,
 
> It defines documents which some may choose to interpret as object
> serializations. What's the problem with that?
> 

There isn't a problem with that at all. How people decide to internally
represent the documents being exchanged is their own business. I can
represent them as pieces of paper, as C structures, as Java objects,
etc. It's about the conceptual model. Web Services are treated as
objects with methods, and WSDL acts as the type description language. We
don't have types in Web Services. We only have messages.

We don't have to go far to see such examples. Look at the Grid community
where an object-based component model is built using WSDL as the means
to describe the interfaces of stateful and transient entities (the Grid
Service Instances). It's the conceptual model and how Web Services are
treated that if of concern. The semantics of what it means to be a
service have been overloaded (statefulness, transient behaviour, SDEs,
etc.) and "requirements" like attributes and inheritance are being
requested from the WSDL group. I've even seen discussions on "service
garbage collection"!

I believe that the term "operation" confuses matters even further and it
should be made clear that there are no operations/methods/procedures but
just message exchanges. I would have preferred for "operation" to be
renamed but there doesn't seem to be wide agreement on this. Perhaps
some language in the WSA document. I believe that we have enough
examples of WSDL's purpose in this world being misunderstood to justify
some appropriate action.

> The WG has taken the decision (repeatedly, IIRC) to stay with the
> term "operation" to describe a message exchange. So the discussion
> on the name is no longer productive.

Apologies for continuing the discussion. I had only seen the decision
for this to get an "issue number". I didn't realise a decision had
already been made not to pursue this any further.

Regards,
.savas.
Received on Monday, 29 September 2003 12:20:07 GMT

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