W3C home > Mailing lists > Public > www-ws-desc@w3.org > February 2004

Re: Issue: should WSDL be able to describe an operation with *anything* in the message?

From: Arthur Ryman <ryman@ca.ibm.com>
Date: Wed, 18 Feb 2004 11:58:47 -0500
To: Jacek Kopecky <jacek.kopecky@systinet.com>
Cc: WS-Description WG <www-ws-desc@w3.org>, www-ws-desc-request@w3.org
Message-ID: <OFFB05E27C.EFD2270E-ON85256E3E.005B9541-85256E3E.005D450A@ca.ibm.com>
Jacek,

IMHO, I don't think it's useful to view a content-based router as a 
service because it is not the ultimate receiver of the message (in the 
SOAP sense) and any client of the service will still have to get the 
"real" WSDL to know what to put into the message.

Wouldn't the more likely case be that the network would be configured to 
route any messages sent to the service through the content-based router? 
i.e. if the client sends a message to the service, the network will 
automatically route it to the content-based router. I assume this just a 
matter of setting up the correct routing tables. The actual URL of the 
content-based router wouldn't be known to the client.

Arthur Ryman,
Rational Desktop Tools Development

phone: 905-413-3077, TL 969-3077
assistant: 905-413-2411, TL 969-2411
fax: 905-413-4920, TL 969-4920
intranet: http://w3.torolab.ibm.com/DEAB/

www-ws-desc-request@w3.org wrote on 02/18/2004 11:06:36 AM:

> 
> Arthur, whether or not a piece of plumbing is a web service (and vice
> versa), that's a different hairy discussion. Let's assume a
> content-based router is, in fact, viewed as a service.
> 
> What's useful about a service description with a single operation that
> accepts anything (and if it has a product, it can be anything again)?
> One probably won't necessarily compile this into a proxy class because
> that is easily superseded by dynamic invocation libraries.
> 
> But when invoking such a service (sending it a message), one wants to
> get the actual address and binding information from somewhere, and WSDL
> is traditionally used for this task. In fact, the tooling should be able
> to handle operations with anything at all in the body quite nicely,
> especially if we accept the likes of Umit's proposal on carrying
> operation identification with messages.
> 
> What do you think?
> 
>                    Jacek Kopecky
> 
>                    Systinet Corporation
>                    http://www.systinet.com/
Received on Wednesday, 18 February 2004 11:58:51 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:15:02 UTC