RE: Proposed resolution to issue 143

 

> -----Original Message-----
> From: Bijan Parsia [mailto:bparsia@isr.umd.edu] 
> Sent: 26 February 2004 12:55
> To: Martin Gudgin
> Cc: WS-Description WG
> Subject: Re: Proposed resolution to issue 143
> 
> Thanks Martin:
> 
> A comment and question inline:
> 
> On Feb 26, 2004, at 7:11 AM, Martin Gudgin wrote:
> 
> [snip]
> > 1.	We agreed on the call that the message attribute always refered
> > to an element declaration in the {element declarations} property of 
> > the definitions component.
> 
> It's my understanding that we are renaming this attribute to 
> "element" 
> to more clearly reflect this fact.

I have no problem with that.

> 
> [snip]
> > 3.	We agreed that references to things that were not elements would
> > require a new collection property, akin to {element 
> declarations} on 
> > the definitions component AND a new attribute in place of 
> the message 
> > attribute to refer to such constructs.
> 
> *AND* a new component property?

Yes, I think so.

> 
> > The spec at[1] has the diffs that cover 1. The changes are 
> in section
> > 2.4 (Message Reference) with identical changes to Section 
> 2.5 (Fault 
> > Reference).
> 
> The diffs, in so far as I can follow them, seem to constrain 
> {message} to element declarations, regardless of what type 
> system the element declarations come from. That's fine, but I 
> just want to check that that is how we're going. In which 
> case, I favor renaming the component to something less 
> generic as well, like "element".
> 
> [snip]
> 
> > I note that 3. is already covered by text in section 3.2:
> >
> > "The extension specification SHOULD, if necessary, define 
> additional 
> > properties of 2.1.1 The Definitions Component to hold the 
> components 
> > of the referenced type system. It is expected that additional 
> > extensibility attributes for Message Reference and Fault Reference 
> > components will also be defined, along with a mechanism for 
> resolving 
> > the values of those attributes to a particular imported type system 
> > component."
> 
> Am I wrong in reading that to say that my extensibility 
> attribute owlClass should populate the current {message} 
> component property with URIs which resolve to components in 
> my new {classes} collection property? If so, that seems to 
> contradict things in section 2.4.

No, I'd expect you to add a new property to the message reference
component. So we should amend the above text to read

"additional properties and extensibility attributes"

Gudge

Received on Thursday, 26 February 2004 11:01:44 UTC