W3C home > Mailing lists > Public > www-ws-desc@w3.org > April 2005

RE: Consistency of WSDL Component property names

From: Jonathan Marsh <jmarsh@microsoft.com>
Date: Wed, 6 Apr 2005 17:46:05 -0700
Message-ID: <7DA77BF2392448449D094BCEF67569A5071774C7@RED-MSG-30.redmond.corp.microsoft.com>
To: "John Kaputin" <KAPUTIN@uk.ibm.com>, <www-ws-desc@w3.org>

This is issue LC107, which we delegated to the editors to change along
the lines you suggest or come back to the WG with further questions.
We're a little behind on our editorial tasks, but expect to implement
this one in the next few weeks.  Thanks for your patience!

[1] http://www.w3.org/2002/ws/desc/4/lc-issues/issues.html#LC107

> -----Original Message-----
> From: www-ws-desc-request@w3.org [mailto:www-ws-desc-request@w3.org]
> On Behalf Of John Kaputin
> Sent: Thursday, March 31, 2005 5:01 AM
> To: www-ws-desc@w3.org
> Subject: RE: Consistency of WSDL Component property names
> 
> 
> I am trying to implement the WSDL 2.0 spec with an API that matches
> the
> WSDL 2.0 component model, but I see that the property name
> inconsistencies
> reported previously still exist in the spec. I couldn't see any
> entries in
> the Issues Lists about this. Can anyone on the WG indicate if/when
> these
> will be corrected?
> 
> There were two issues:
> 1. the word 'reference' was incorrectly used in some property names to
> refer to components that were not XXXReference components
> 2. whether to use short-hand names like {fault} and {faults} or full
> descriptive names like {interface fault} and {binding faults}
> 
> I am most concerned with issue 1. Three property name changes are
> required:
> 
> FaultReference {fault reference}  becomes  {fault} or {interface
> fault}
> (because the property refers to an InterfaceFault component)
> 
> BindingFault {fault reference}  becomes  {fault} or {interface fault}
> (because the property refers to an InterfaceFault component)
> 
> BindingOperation {operation reference}  becomes  {operation} or
> {interface
> operation}
> (because the property refers to an InterfaceOperation component)
> 
> Note, the following uses of 'reference' are correct:
> 
> InterfaceOperation {fault references}  .... a set of FaultReference
> InterfaceOperation {message references} .... a set of MessageReference
> BindingMessageReference {message reference} .... a MessageReference
> BindingOperation {message references} .... a set of Binding Message
> Reference (1)
> BindingOperation {fault references} .... a set of
> BindingFaultReference (2)
> BindingFaultReference {fault reference} .... a FaultReference
> 
> (1) so maybe {binding message references} could be used?
> (2) so maybe {binding fault references}?
> 
> John Kaputin
> Hursley Laboratory
> IBM UK
> ----- Forwarded by John Kaputin/UK/IBM on 31/03/2005 12:02 -----
> 
>              Asir Vedamuthu
>              <asirv@webmetho
>              ds.com>
> To
>                                      John Kaputin/UK/IBM@IBMGB,
>              11/02/2005              www-ws-desc@w3.org
>              22:22
> cc
> 
> 
> Subject
>                                      RE: Consistency of WSDL Component
>                                      property names
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> +1, continuing along these lines, I request the following changes ...
> 
> Binding Operation.{operation reference} =>
>     Binding Operation.{interface operation}
> 
> Binding Operation.{message references} =>
>     Binding Operation.{binding message references}
> 
> Taking LC55 [1] into account ...
> 
> Binding Fault Reference.{fault reference} =>
>     Binding Fault Reference.{interface fault reference}
> 
> Binding Message Reference.{message reference} =>
>     Binding Message Reference.{interface message reference}
> 
> 
> I request the WG to consider the following ...
> 
> Similar to (Interface Operation, Binding Operation), (Interface Fault,
> Binding Fault), ...
> 
> Fault Reference => Interface Fault Reference
> Message Reference => Interface Message Reference
> 
> That leads to ...
> 
> Interface Operation.{fault references} =>
>     Interface Operation.{interface fault references}
> 
> Interface Operation.{message references} =>
>     Interface Operation.{interface message references}
> 
> [1] http://www.w3.org/2002/ws/desc/4/lc-issues/issues.html#LC55
> 
> Regards,
> Asir S Vedamuthu
> asirv at webmethods dot com
> http://www.webmethods.com/
> 
> -----Original Message-----
> From: www-ws-desc-request@w3.org [mailto:www-ws-desc-request@w3.org]
> On
> Behalf Of John Kaputin
> Sent: Thursday, February 10, 2005 8:04 AM
> To: www-ws-desc@w3.org
> Subject: Consistency of WSDL Component property names
> 
> 
> 
> 
> 
> 
> I'd like to suggest some improvements in the consistency of property
> names
> in the Component Model (WSDL 2.0 Part 1 spec, Section 2 Component
> Model)
> 
> ElementDeclaration is referred to by properties in various components:
> 
> Description has property {element declarations}  - a set of
> ElementDeclaration
> InterfaceFault has property {element}                  - an
> ElementDeclaration
> MessageReference has property {element}         - an
> ElementDeclaration
> 
> For clarity, could same name be used for properties that refer to the
> same
> type of component (with adjustments for plural or singular):
> Description {elements}
> InterfaceFault {element}
> MessageReference {element}
> 
> ==================
> 
> There is a similar inconsistency with the names of fault properties:
> 
> Interface {faults} - a set of InterfaceFault
> InterfaceOperation {fault references} - a set of FaultReference
> FaultReference {fault reference} - an InterfaceFault
> BindingFault {fault reference} - an InterfaceFault
> 
> The use of {fault references} for InterfaceOperation makes sense, but
> its
> use in FaultReference and BindingFault is confusing. For example, a
> FaultReference {fault reference} refers to an InterfaceFault that must
> be a
> member of the parent Interface {faults} so why not use the same
> property
> name for both?
> 
> In this example, the {fault reference} property in FaultReference and
> BindingFault could be simply {fault}, thus:
> FaultReference {fault} - an InterfaceFault
> BindingFault {fault} - an InterfaceFault
> 
> ==========================
> 
> Perhaps property names could be made not only consistent but more
> descriptive by basing them on the name of the Component they refer to.
> This
> would facilitate the creation of APIs based closely on the WSDL
> Component
> Model that are more descriptive (eg: the getter/setter methods for
> properties).
> 
> Thus...
> 
> Description {element declarations} - a set of ElementDeclaration
> InterfaceFault {element declaration} - an ElementDeclaration
> MessageReference {element declaration} - an ElementDeclaration
> 
> and...
> 
> Interface {interface faults} - a set of InterfaceFault
> InterfaceOperation {fault references} - a set of FaultReference
> FaultReference {interface fault} - an InterfaceFault
> BindingFault {interface fault} - an InterfaceFault
> 
> This would also clarify the use of operations and faults across
> Interfaces
> and Bindings:
> 
> Interface {fault} - an InterfaceFault
> Binding {fault} - a BindingFault
> 
> could become...
> 
> Interface {interface fault} - an InterfaceFault
> Binding {binding fault} - a BindingFault
> 
> And....
> 
> Interface {operations} - a set of InterfaceOperation
> Binding {operations} - a set of BindingOperation
> 
> could become...
> 
> Interface {interface operations} - a set of InterfaceOperation
> Binding {binding operations} - a set of BindingOperation
> 
> 
> regards,
> John Kaputin
> Hursley Laboratory
> IBM UK Ltd
> 
> 
Received on Thursday, 7 April 2005 00:46:24 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:35 GMT