W3C home > Mailing lists > Public > www-xml-schema-comments@w3.org > July to September 2007

Re: [Bug 4886] proposal for alternative language for defining derived datatypes

From: Hedenus, Michael (EXT) <Michael.Hedenus.ext@siemensvdo.com>
Date: Fri, 28 Sep 2007 19:19:31 +0000
To: <www-xml-schema-comments@w3.org>
Message-Id: <02C568DE-707F-4728-8E7F-12BD1CED2F41@siemensvdo.com>
Cc: "Dave Peterson" <davep@iit.edu>
Ladies and Gentlemen,

thank you very much for regarding my proposal.

(1)
The original intention was a proposal for expressing simple types  
only (not even unions) as URIs, because expressing a complex type as  
an URI would be overdoing (otherwise we could say why not expressing  
a whole Schema as an URI). So I still propose to define a mechanism  
how to add facets to the URI reference of a type.

Consider the following code fragment as an example, where an int is  
restricted so that its value must be between 10 and 100 inclusive.

<rdf:Description about="...">
   <my:property datatype="http://www.w3.org/2001/XMLSchema#int 
[maxInclusive='100',minInclusive='10']">42</my:property>
</rdf:Description>

Here facets are just added in squared brackets and separated by comma.

(2)
If you want to refer to a complex type (or to any type definition),  
one could do it using the id attribute.
I faced this kind of problem while getting along with "typed XML  
literals" in RDF (although this is not considered by the RDF  
specification I see no contradiction):

   Schema file located at http://mycompany/schema.xsd:
   --------------------------------------------------------

   <xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
     <xs:complexType id="mytype" name="mytype">
    .. <xs:element name="elem1">...
     <xs:complexType>
   </xs:schema>


   RDF file:
   ---------

   <rdf:Description about="...">
     <my:property rdf:parseType="Literal" rdf:datatype="http:// 
mycompany/schema.xsd#mytype">
       <elem1>...</elem1>
     </my:property>
   </rdf:Description>


As you see, especially RDF could take advantage of such a mechanism.  
But to get this working one has to specify that it is legal to refer  
to a type via its attribute. This is not a RDF issue because RDF does  
not make any assumption about the data type reference itself, only  
that it must be an URI reference.
So I propose to define such a referencing mechanism to a type  
definition within a Schema via its id attribute.

(3)
In my humble opinion the whole issue is important enough not to be  
closed.
If agreed, I offer to prepare a formal specification.

PS:
  Imagine the RDF parser detecting the type, auto-creating a data- 
binding and reading the data into a data-object like a JavaBean...

Mit freundlichen Grüßen

Dr. Michael Hedenus
Pixel GmbH

Externer Dienstleister der Siemens VDO Automotive AG
SV P EG 1
Siemensstrasse 12
93055 Regensburg, Germany
Tel.: +49 (941) 790-6362
Fax: +49 (941) 79013-6362
mailto:Michael.Hedenus.ext@siemensvdo.com

PIXEL GmbH, Lochhammer Schlag 17, 82166 Gräfelfing bei München
Geschäftsführer: Jehanger Grami, Simon Ashdown, Manfred Kügler
Sitz der Gesellschaft: Gräfelfing, Amtsgericht: München HRB 86287

Wichtiger Hinweis: Diese E-Mail und etwaige Anlagen enthalten  
firmenvertrauliche Informationen. Sollten Sie diese E-Mail irrtümlich  
erhalten haben, benachrichtigen Sie uns bitte durch Antwort-Mail und  
löschen Sie diese E-Mail nebst Anlagen von Ihrem System. Vielen Dank.
Received on Friday, 28 September 2007 19:30:10 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:50:07 UTC