W3C home > Mailing lists > Public > www-ws-desc@w3.org > April 2002

RE: R067 and R967a: extensibility

From: Sedukhin, Igor <Igor.Sedukhin@ca.com>
Date: Fri, 5 Apr 2002 13:11:30 -0500
Message-ID: <849C1D32E4C7924F854D8A0356C72A9E0307EA2F@usilms08.ca.com>
To: Jeffrey Schlimmer <jeffsch@windows.microsoft.com>, www-ws-desc@w3.org
Jeff, I agree with your wording of R067. I also agree to the requirement itself.
Now, my additional point was to have a *new* requirement, in addition to R067. I call it R067a.
The new requirement would use the extensibility mechanism provided by R067 and define extra constructs in the new spec to *hold* the application-specific metadata in a flexible manner (XML).
R067 leaves it open for extensions.
R067a defines one of the extensions. (very similar to the current bindings spec).
The requirements have different intended audience. I'd very much like to see that new WSDL spec provides for both.
Even if syntactically these requirements coincide in the spec, two requirements would at least ask us to provide separate examples for the use cases. And that means real-world implementations will have something to relate to as part of a standard, not proprietary extension.
-- Igor Sedukhin .. (Igor.Sedukhin@ca.com) 
-- (631) 342-4325 .. 1 CA Plaza, Islandia, NY 11788 

-----Original Message-----
From: Jeffrey Schlimmer [mailto:jeffsch@windows.microsoft.com] 
Sent: Friday, April 05, 2002 12:47 PM
To: www-ws-desc@w3.org
Subject: RE: R067 and R967a: extensibility

Igor, I agree that the constructions defined in WSDL should be extensible; R067 enumerates the constructions defined in WSDL, but the intent might be clearer if the requirement referred to constructions as a set. In the next draft, I reworded R067 to the following:


R067 [?Modify?, Draft, Must, WG] The WG will ensure that there are adequate points of extensibility in ?each construction defined by WSDL (e.g., Operation, Interface, ...).? (Last discussed 4 Apr 2002.)


I also agree that any extensions should themselves be appropriately extensible ("ability to define extensible"), but shouldn't the characteristics of an extension be the sovereignty of the extension, not WSDL?




-----Original Message-----
From: Sedukhin, Igor [mailto:Igor.Sedukhin@ca.com] 
Sent: Thursday, April 04, 2002 3:30 PM
To: www-ws-desc@w3.org
Subject: R067 and R967a: extensibility


As a followup to the conf.call discussion: 

R067: would refer to "the extensibility of the definitions mandated by the WSDL specification" 
R067a: would refer to "the ability to define extensible (application-specific) metadata around constructs mandated by the WSDL specification"

Here is the semantic difference: 

Results of R067 would allow one to define a WSDL-derived specification with basic constructs defined by WSDL itself and extra constructs defined by extensions. For instance, bindings as defined now fall into this category. We just have to cover other constructs similarly.

Here is an example where elements in bold define extensions of the definitions. 
<message name="nmtoken"> *
<-- extensibility element (1) --> *
<part name="nmtoken" element="qname"? type="qname"> *
<-- extensibility element (2) --> *

Results of R067a would, for instance, allow one to include a link to a UML diagram of a use case as a metadata element for a definition of a service.

Here is the actual example. 
<extend xmlns:myExt="...">
<myExt:useCase id="purchaseOrder" UML="http://mycorp/usecases/PO.vsd <http://mycorp/usecases/PO.vsd> "/>
<service extend="purchaseOrder" ... 
<service ... >
<extend name="purchaseOrder" xmlns:myExt="...">
<myExt:useCase id="purchaseOrder" UML="http://mycorp/usecases/PO.vsd <http://mycorp/usecases/PO.vsd> "/>

And this is something I should be able to use as follows (UC0026) 
NodeList ext = client.getService().getExtend(); 
Element useCase = (Element)NodeList.item(0); 
URL linkToUML = new URL(useCase.getAttribute("UML")); 


-- Igor Sedukhin .. (Igor.Sedukhin@ca.com) 
-- (631) 342-4325 .. 1 CA Plaza, Islandia, NY 11788 

Received on Friday, 5 April 2002 13:12:11 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:54:37 UTC