RE: Completing Part 1 Appendix C: URI References for WSDL constructs

Thanks for your comments.  We resolved them as detailed below.  If you
don't respond by October 1, we'll assume you accept these resolutions.

> -----Original Message-----
> From: [mailto:public-ws-desc-
>] On Behalf Of Hugo Haas
> Sent: Tuesday, September 14, 2004 8:58 AM
> To:
> Subject: Completing Part 1 Appendix C: URI References for WSDL
> Appendix C in Part 1 is not very visible despite being useful, and is
> incomplete.
> First, I believe that we should point to Appendix C every time we talk
> about not being able to a component by a simple QName, such as in
> sections 2.3.1, 2.4.1 and 2.14.1, but there may be others that I have
> missed.

The WG agreed to add links to the Component Designator definition to
sections of specification where we note that QName is insufficient to
identify a component.


> Second, section C.2 is lacking references for the following components
> in order for us to fulfill R120, "The description language MUST ensure
> that all conceptual elements in the description of Messages are
> addressable by a URI reference"; I am proposing some solutions:
> - message reference: messageref(interface/operation/direction/label);
> - fault reference: faultref(interface/operation/direction/label/ref);
> - binding fault: bindingfault(binding/ref);
> - binding operation: bindingoperation(binding/ref);
> - binding message reference:
>   bindingmessageref(binding/operation/direction/label).

The WG accepts the proposal in general (though the syntax may change
slightly to remain consistent with other schemes), and also to include
F&P schemes.


> Third, some of the construct names in the first column are not
> corresponding exactly to components names:
> - s/operation/interface operation/;
> - s/fault/interface fault/.

The WG agreed to accept the proposal, and will change the names in table
C-1 to match component names.


> Finally, I find the the x / y convention hard to read. What about
> something like:
>   endpoint: _endpoint_ being the {name} property of endpoint and
>   _service_ being the {name} property of parent service:
>   endpoint(_service_/_endpoint_)
> with _foo_ being typographically different, e.g. in italic?

The WG agreed with this suggestion, and will also rework the table with
other readability improvements as the new components are added.


Received on Tuesday, 21 September 2004 18:21:55 UTC