W3C home > Mailing lists > Public > www-ws-desc@w3.org > February 2004

Re: Message Reference, Message|element encore

From: Bijan Parsia <bparsia@isr.umd.edu>
Date: Thu, 12 Feb 2004 11:50:24 -0500
Message-Id: <8DA975FA-5D7B-11D8-BB61-0003939E0B44@isr.umd.edu>
To: www-ws-desc@w3.org

On Feb 11, 2004, at 5:44 PM, Bijan Parsia wrote:
[snip]
> All this suggests (to me) that having to add an attribute for each 
> type system is, well, annoying :) Why not have a pair of component 
> properties, {typeSystem} and {type}. And better, let there be two 
> attributes in the XML as well. For XML Schema element declarations, we 
> can make that omitting the type system attribute defaulst to XML 
> Schema element declarations.

So, Arthur pointed out that the type of the attribute is qname which, 
for some type systems (like OWL) might be annoying (since OWL names 
classes with URIs). This would be a reason to avoid the generic 
attributes (or assert that every extended type system supply qname 
identifiers or a way of constructing a qname from their native 
identifiers). I wonder how this affects the component model...can the 
type of the {message} property be a pair ({type system, type system 
specific identifer})?

If every type system must add a component property as well, then I 
think it's a good idea to make the {message} component more specific 
(i.e., {element}). But then the inconsistency of the text should be 
resolved in favor of making {element} *only* take its values from the 
element attribute.

Cheers,
Bijan Parsia.
Received on Thursday, 12 February 2004 11:50:28 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:15:02 UTC