W3C home > Mailing lists > Public > public-swbp-wg@w3.org > November 2005

Re: [PORT] suggestion re Collections and SKOS/OWL - SKOS version of Wine Onto?

From: Aldo Gangemi <aldo.gangemi@istc.cnr.it>
Date: Fri, 04 Nov 2005 18:53:01 +0100
To: Dan Brickley <danbri@w3.org>, public-swbp-wg@w3.org
Message-id: <p06210207bf914845fc5e@[]>
At 11:10 -0500 4-11-2005, Dan Brickley wrote:
>I don't know the best URI to cite, but
>...has a handy overview of the Wine Ontology. I think it
>would be very interesting to try a SKOS representation of this,
>to explore how a thesaurus (eg. faceted) approach looks.

My idea at the discussion was that the semantics of a node label is 
similar to a complete definition of an owl class, e.g. in the wine 

  intersection of:
  on property: http://www.w3.org/TR/2003/CR-owl-guide-20030818/wine#hasColor
  has value: http://www.w3.org/TR/2003/CR-owl-guide-20030818/wine#White

which just says that a wine:WhiteWine is a wine that wine:hasColor 
wine:White. In a thesaurus, this would  be represented as WhiteWine 
being an NT of Wine, eventually having an RT semantic relation to 
Color, or even to White. Of course in thesauri, no constraint is 
given on the semantics of NT or RT.
Now suppose that that wine thesaurus wants to introduce a node label 
for "wines-by-color", which would be a "collection" having as 
"member" WhiteWine, in your current solution for node labels.
But if we go back to owl, the semantics of the wine-by-color node 
label would be most likely the following:

  intersection of:
  on property: http://www.w3.org/TR/2003/CR-owl-guide-20030818/wine#hasColor
  cardinality: 1

i.e. that a wine-by-color is a wine that has a cardinality=1 on the 
property hasColor. A DL reasoner will easily infer that WhiteWine is 
a subClassOf WineByColor.
So far for the semantic issue. And that's why I dislike the 
disjointness between concepts and collections in SKOS.

As far as the pragmatic issue is concerned, i.e. that node labels are 
not used for annotating, it can be solved in several ways (I suggest 
to add an annotation property to a node-label-derived class), but not 
within object-level OWL(DL), and motivately so: usage of metadata 
should not be part of their set-theoretic semantics.

On the other hand, SKOS domain is not similar to the Wine domain: it 
is about concepts, not wines, then one may claim that the usage of 
metadata concerns exactly the semantics of concepts. And here we are 
back to the same alternative found in th WN TF: are thesaurus 
(wordnet) concepts (synsets) to be treated as instances, or migrated 
to classes?
SKOS does the same as the WN datamodel in OWL, and treats concepts as 
instances. Therefore, we may be tempted to do something weird to 
node-label-derived instances.
What I suggest is not btw to make a partition concepts/collections, 
but to introduce a new property that allows "node-label" assertions 
to some instances of the "concept" class, e.g.:

Class(Concept partial
  restriction(hasNodeLabel allValuesFrom(xsd:boolean)))

  value(hasNodeLabel, T))

and eventually introduce a special subClass for node labels:

Class(NodeLabel complete
   restriction(hasNodeLabel oneOf(T)))

If the problem is that thesauri experts refuse to call node labels 
"concepts", just create a more general class, e.g. "descriptor", and 
make both concepts and node labels subclasses of it:

Class(NodeLabel complete
   restriction(hasNodeLabel oneOf(T)))

Class(Concept complete
   restriction(hasNodeLabel oneOf(F)))


Aldo Gangemi
Research Scientist
Laboratory for Applied Ontology
Institute for Cognitive Sciences and Technology
National Research Council (ISTC-CNR)
Via Nomentana 56, 00161, Roma, Italy
Tel: +390644161535
Fax: +390644161513
Received on Friday, 4 November 2005 22:25:31 UTC

This archive was generated by hypermail 2.4.0 : Friday, 27 January 2023 01:58:25 UTC