W3C home > Mailing lists > Public > www-rdf-comments@w3.org > July to September 1998

Typed collections and cardinalities

From: Didier Villevalois <Didier.Villevalois@inria.fr>
Date: Wed, 02 Sep 1998 12:03:17 +0200
Message-ID: <35ED17E5.BE7D484B@inria.fr>
To: www-rdf-comments@w3.org
Hello,

These are general comments about the two last public drafts of RDF:
 - RDF Model and Syntax Spec. (August 19, 1998)
 - RDF Schema Spec. (August 14, 1998)

I've allready wrote before some of these comments without any answers.
So i don't know if those are stupid, boring or not, or if i
misunderstood RDF's goal or features.

Typing elements of collections
==============================

As we can type the range of a property, typing collections content with
a property such as the instanceOf property would be a very practical
stuff. There is actually no mean to do that.

Such a constraining property for collections could be seen as an
constraining distributed instanceOf above elements of these collections.

As a resulting fact this would permit a simplifying writing of an RDF
instance. This could be done by optionnaly ommiting the rdf:li tag
without confusion between embeded elements of the collection and
properties for this collection.

For instance we could write something like:
 ...
 <rdf:Bag>
  <rdf:elementsInstanceOf resource="s:Member"/>
  <s:Member>Ora</Member> <!-- An embeded element -->
  <myMeta:writenOn>09-02-1998</myMeta:writenOn> <!-- A property -->
  <rdf:li resource="#Joe"/> <!-- A refered element -->
  <s:Chairman>Ralph</s:Chairman>
  <!-- An embeded element whose class is a sub-class of the Member class
-->
 </rdf:Bag>
 ...

which would be quite equivalent to:
 ...
 <rdf:Bag ID="people">
  <s:Member>Ora</Member> <!-- An embeded element -->
  <myMeta:writenOn>09-02-1998</myMeta:writenOn> <!-- A property -->
  <rdf:li resource="#Joe"/> <!-- A refered element -->
  <s:Chairman>Ralph</s:Chairman>
  <!-- An embeded element whose class is a sub-class of the Member class
-->
 </rdf:Bag>

 <rdf:Description aboutEach="#people">
  <rdf:instanceOf resource="s:Member"/>
 </rdf:Description>
 ...

I think this is different because we cannot define a schema in the
second way to say that a class that is instance of a collection to have
typed elements. This could be then possible with the first way.

With such a stuff, any actual XML/DTD document instance may be
interpreted in the RDF model without any changes to this instance.

Cardinalities
=============

Isn't there anymore any mechanism to express cardinality constrains ?

Thanks for your answer.
Didier Villevalois.

--
Didier Villevalois - Didier.Villevalois@inria.fr
Rodin Project - INRIA, Rocquencourt, B.P. 105, 78153 Le Chesnay, France
Work Phone: +33 1 39 63 56 18 - Fax: +33 1 39 63 51 93
Received on Wednesday, 2 September 1998 05:01:25 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 21 September 2012 14:16:26 GMT