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

RE: What does WSDL describe?

From: Savas Parastatidis <Savas.Parastatidis@newcastle.ac.uk>
Date: Fri, 24 Oct 2003 00:01:01 +0100
Message-ID: <BC28A9E979C56C44BCBC2DED313A4470021C388E@bond.ncl.ac.uk>
To: <www-ws-desc@w3.org>
Cc: "Jim Webber" <jim.webber@arjuna.com>, "Mark Baker" <distobj@acm.org>

Mark,
 
> 
> I'm going to hazard asking this seemingly dumb question, because
> frankly, I just don't seem to know the answer anymore (after reading
> some of the answers to Umit's question, but also some of Jim
> Webber's messages earlier).
> 

I can't talk for Jim Webber but from private discussions with him I
think we have a common understanding on this (he may not see this for
some time since he has chosen diving in the Red Sea and hiking around
mount Everest for the next 2 1/2 months, instead of following the WSDL
group discussions... some people can't get their priorities right :-).

> What does WSDL 1.2 describe?
> 

To my mind, a WSDL document defines the messages and their structure
that a Web Service is prepared to accept/send. It is effectively a
contract between the service and its consumers about the message
exchange patterns (MEPs) in which the web service is prepared to
participate and the structure of the messages that make up those MEPs.

Furthermore, the binding part of a WSDL document defines the network
communication means by which the messages can be sent to/received by
that web service. The service part identifies the endpoint for the
communication.

I wouldn't call the above description as a "document endpoint" but
effectively we are talking about exchange of documents (the messages).

I don't think that a WSDL document says anything about the semantics of
the message exchanges or the semantics of the message contents. For
example, we don't know whether an element in one of the messages of type
xsd:string is the name of a gene sequence or a person's name. We don't
know, and I don't think we should assume, whether a WSDL "operation"
called "buyApple" will result in an apple or the company Apple being
bought. All a WSDL document defines is the structure of the messages
exchanged.

Do we need more information about what the exchange of messages mean? We
have to go the semantics people and look at what specifications like
DAML-S can provide us.

Web services are all about machine-to-machine communication. WSDL plus a
schema language provide the means through which we can automate the
validation of the documents being exchanged. The way WSDL has been used
to date (i.e., mostly to automate proxy-generation through which web
services are seen as objects with methods) has made people think that
WSDL is yet another IDL.


That's my take on your question. All the above are the reason Jim and I
suggested that "operation" be renamed to "messageExchange". As someone
mentioned in an earlier message (or, was it in the minutes of a
teleconference?), perhaps renaming "interface" should also be
considered... "messageExchangeGroup" perhaps?

I am looking forward to seeing what other people think about all these.
I may be way off on my understanding of WSDL.

Best regards,
.savas.
Received on Thursday, 23 October 2003 19:01:09 GMT

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