RE: simple case of IRIs for Components in WSDL 2.0

IMO trying to distinguish the form of the identifier depending upon
whether it is dereferenced to application/wsdl+xml or not would be
highly suspect.  I agree with Dan here (IIUC).

 

The side-effect of adopting Dan's proposal is that we would have a +xml
media type that does not conform to XPointer.  That decision is within
our power as registrars of the media type, but seems to me to be a
pretty horrible idea, especially since the suggested syntax would be
indistinguishable from that of XPointer, but with a different semantic.
In addition, our current syntax allows you to point to either a
component, or an XML element.  By conflating the syntax, we'd lose the
latter.  The benefits of breaking with XPointer, making identifiers a
little more aesthetically pleasing in some cases while adding complexity
to the overall identification algorithm, seem marginal at best.

 

________________________________

From: Arthur Ryman [mailto:ryman@ca.ibm.com] 
Sent: Thursday, September 15, 2005 2:36 PM
To: Dan Connolly
Cc: Bijan Parsia; Henry S. Thompson; Jonathan Marsh;
public-ws-desc-comments@w3.org; public-ws-desc-comments-request@w3.org
Subject: RE: simple case of IRIs for Components in WSDL 2.0

 


Dan, 

> In case 2 we have more leeway because the URI reference is being used
> as a component identifier and the namespace is not being referenced.
> 
I don't understand what you mean by that.


I meant to say that the namespace is not being dereferenced. In this
case we are using the namespace simply as a name and are forming
identifiers based on it using the URI reference syntax. 

Even if the namespace might be dereferencible, it might not be a WSDL
document. It might be a RDDL document or something else. The point is
that we are  not assuming anything about a media type. We are just
forming identifiers. 

Arthur Ryman,
IBM Software Group, Rational Division

blog: http://ryman.eclipsedevelopersjournal.com/
phone: +1-905-413-3077, TL 969-3077
assistant: +1-905-413-2411, TL 969-2411
fax: +1-905-413-4920, TL 969-4920
mobile: +1-416-939-5063, text: 4169395063@fido.ca 

Dan Connolly <connolly@w3.org> 
Sent by: public-ws-desc-comments-request@w3.org 

09/15/2005 05:09 PM 

To

Arthur Ryman/Toronto/IBM@IBMCA 

cc

public-ws-desc-comments-request@w3.org, public-ws-desc-comments@w3.org,
Jonathan Marsh <jmarsh@microsoft.com>, "Henry S. Thompson"
<ht@inf.ed.ac.uk>, Bijan Parsia <bparsia@isr.umd.edu> 

Subject

RE: simple case of IRIs for Components in WSDL 2.0

 

 

 





On Thu, 2005-09-15 at 16:51 -0400, Arthur Ryman wrote:
> 
> Dan, 
> 
> There are two cases we need to consider when forming URI references. 
> 
> 1. The URI deferences to a WSDL document whose media type is
> application/wsdl+xml 
> 2. The URI is a WSDL namespace and we are forming identifiers for
> components. 
> 
> I'm not an XPointer lawyer, but in case 1 I don't see how we can
> change the meaning of the bare names. Doesn't that violate the
> XPointer spec?
>
> Is application/wsdl+xml allowed to redefine the meaning of XPointer
> for application/xml? 

It's not a matter of redefining.

You can choose whether and to what extent the media type definition
of application/wsdl+xml inherits from XPointer.

I don't have any requirements for you to use XPointer at all.

I do have a requirement that you don't use XPointer's barename
definition.

> In case 2 we have more leeway because the URI reference is being used
> as a component identifier and the namespace is not being referenced.
> 
I don't understand what you mean by that.

Whatever you get by dereferencing a URI should agree with other
specifications about that URI.
> 
>  In this case would could define bare names to mean whatever we want.
> So we could define a bare name to identify the WSDL component with
> that local name, assuming it was unique. 
> 
-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E