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

Dicussion of alt schema languages

From: Bijan Parsia <bparsia@isr.umd.edu>
Date: Thu, 26 May 2005 11:33:49 -0400
Message-Id: <81adcd51dc40a4c7246f338d6637e8d7@isr.umd.edu>
To: www-ws-desc@w3.org

http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl20/ 
altschemalangs.html

Here are the issues I see still existing with regard to this document.

(Hmm. I notice it's not quite up to date, e.g., wrt the Editors list  
and an intro that I wrote. I'll try to resolve my cvs issues.)

1) There are no examples (or discussion) of using a non-XML type system  
extension.
2) There is no discussion of how to handle multiple definitions of the  
same type/element. E.g., a relax ng and a XML schema characterization  
of the same element. (Must they be (provably) equivalent? Which can  
you/must you use?)
3) There is no discussion of how to handle operations with types that  
span multiple schema langauges. E.g., where we have a union output type  
where each alternative has a different element and is specified by a  
distinct schema language. (Real use case is from the Data Access  
working group where in they which to say that the results of of a query  
can be either the xml results format (<results>), specified by a XML  
Schema, or an RDF document (<rdf:RDF>) specified by a Relax NG schema).

I suggest minimality in resolving them. Perhaps just *mentioning* that  
2 and 3 have to be resolved by an extension. OTOH, we might want to say  
something about how element attributes work, i.e., to fix how they  
behave in the presence of extensions.

Cheers,
Bijan.
Received on Thursday, 26 May 2005 15:33:51 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:36 GMT