Guide review

Review copy dated 18 december

Scope of review:
http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0320.html


IMO the best of our documents.

Overall comments (in my scope)

The discussion of imports is insufficient, since the interaction with levels 
of OWL is complicated.

There should be additional sections concerning the relationship with RDF/XML 
documents particularly clarifying that the ise of bnodes in OWL DL and OWL 
Lite is restricted and that the classes and some properties must be declared 
appropriately.


Introduction
=========

"There is nothing preventing ... exclusively individuals"

Misleading, since, in owl Lite and OWL DL, and triple

<x> rdf:type <eg:foo> .

must be accompanied by a triple

<eg:foo> rdf:type owl:Class .

Suggest replace "exclusively" with "nearly exclusively".

The Species of OWL
================
Suggest add a paragraph indicating that when creating OWL documents in RDF/XML 
by hand significant additional care is needed to ensure that it meets the 
syntactic requirements of OWL DL, and much more with respect to OWL Lite.


Namespaces
===========

I note that the cardinality examples have not been updated to reflect the 
datatyping decisions.
I suggest the extension of the DTD internal subset as described in

http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0254.html


Ontology Headers
===============

Discussion of imports does not deal with the limitations of import in its 
inability to meet the syntactic requirements associated with OWL Lite and OWL 
DL. Specifically the fact that the OWL Lite file:

<x> rdf:type <eg:foo> .
<eg:foo> rdf:type owl:Class .

cannot be split into two files with an imports.


The comment about Dublin Core is misleading for OWL DL and OWL Lite see my 
comments in my Feature Synopsis Review.


Simple Properties
==============

With the  first example, it appears to be in OWL Lite, but it requires the 
supporting definitions
<owl:Class rdf:ID="Wine"/>
<owl:Class rdf:ID="WineGrape"/>
(without these it is in OWL Full).

It is not clear to me how/where best to express this. There is quite a lot of 
other implicit stuff concerning this example - and I don't want to highlight 
this single issue over and above say xml namespaces.

Similarly in the example with the minCardinality there is a need to decide 
whether to explicitly add the rdf:datatype attributes or whether to use 
default attributes as I have suggested. In the latter case there is a 
presentational issue about how/where to explain it.
 

Properties and Datatypes
====================
I thought we had dropped the user defined types, in which case this whole 
section needs rework (or deletion).
I note that user defined types are in the abstract syntax document ....



Complex Classes
Intersection

The following very similar class (semantically identical) is in OWL Lite.
I wonder if you can manage to explain it? (I couldn't).

<owl:Class rdf:ID="WhiteWine">
  <owl:sameClassAs>
   <rdf:Description>
  <owl:intersectionOf rdf:parseType="Collection">
    <owl:Class rdf:about="#Wine" />
    <owl:Restriction>
      <owl:onProperty rdf:resource="#hasColor" />
      <owl:hasValue rdf:resource="#White" />
    </owl:Restriction>
  </owl:intersectionOf>
  </rdf:Description>
  </owl:sameClassAs>
</owl:Class>

Enumerated Classes (oneof)

You could also give an example of a datatype range from the abstract syntax.

<dataRange> ::= oneOf({<typedDataLiteral>} )

(difficult to do because the parseType="Collection" syntax is not usable).


References

Any chance of referencing the more recent RDF WDs?

Received on Tuesday, 31 December 2002 12:16:56 UTC