W3C home > Mailing lists > Public > semantic-web@w3.org > January 2011

Re: About Punning

From: Antoine Zimmermann <antoine.zimmermann@insa-lyon.fr>
Date: Mon, 10 Jan 2011 17:31:30 +0100
Message-ID: <4D2B3462.8060302@insa-lyon.fr>
To: Cristiano Longo <longo@dmi.unict.it>
CC: semantic-web@w3.org
In OWL 2 DL, punning won't make a class a property or vice versa. What 
punning allows you to do is to use the same name (same IRI) to identify 
two *distinct* things.

So, if you define the following:
:Marriage a owl:Class, owl:ObjectProperty .

You actually define two different terms (a class and a property) which 
have, a priori, nothing in common. However, they share the same name, 
which can be interesting in some application.

Depending on the situation, you may want to write the followin:

:Marriage a owl:Class .
:joesMarriage a :Marriage ;
     :date "2008-08-25"^^xsd:date .
     :hasband :joe ;
     :wife :mary .

:Marriage a owl:ObjectProperty .
:joe :Marriage :mary .

This looks very neat, but in fact, :Marriage(the class) and 
:Marriage(the property) are really unrelated and no constraint 
whatsoever on one of them would influence the other.
For instance, you could as well say:

:Marriage a owl:Class ;
     owl:equivalentClass owl:Nothing .
:Marriage a owl:ObjectProperty .
owl:topObjectProperty rdfs:subClassOf :Marriage .

Which says that :Marriage(the class) is empty (i.e., has no instance) 
and that :Marriage(the property) relates everything to everything. It is 
consistent.

In fact, I find punning mostly useful when a class name is used as an 
individual name, e.g.:

:UserGroup a owl:Class .
:Friends a :UserGroup, owl:Class .
:CloseFriends a :UserGroup;
     rdfs:subClassOf :Friends .
:joe a :CloseFriend .

In this case, you emulate group membership using rdf:type, and subgroup 
relation using rdfs:subClassOf, which is much more powerful than having 
a custom :member property or :subGroup property (note that the subset 
relation cannot be modelled in OWL 2 DL, unless one is using 
rdfs:subClassOf, which has exactly the semantics of subset).


Le 08/01/2011 12:08, Cristiano Longo a écrit :
> Does Punning in OWL 2 solves all the issues related to meta-modelling
> (and reasoning against meta-classes and meta-propeties), or there are
> yet some open issues? In particular, in OWL 2 DL can I
>
> (a) define an hybrid-object which contains both objects and pairs?
> (b) can I relate two properties?
>
> Thank you in advance,
> Cristiano Longo
>

Regards,
-- 
Antoine Zimmermann
Researcher at:
Laboratoire d'InfoRmatique en Image et Systèmes d'information
Database Group
7 Avenue Jean Capelle
69621 Villeurbanne Cedex
France
Lecturer at:
Institut National des Sciences Appliquées de Lyon
20 Avenue Albert Einstein
69621 Villeurbanne Cedex
France
antoine.zimmermann@insa-lyon.fr
http://zimmer.aprilfoolsreview.com/
Received on Monday, 10 January 2011 16:32:06 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:48:20 UTC