- From: Josh Sled <jsled@asynchronous.org>
- Date: Sat, 10 Jul 2004 10:52:34 -0400
- To: Laurian Gridinoc <laurian@gmail.com>, Richard Newman <r.newman@reading.ac.uk>
- Cc: www-rdf-interest@w3.org
On Fri, 2004-07-09 at 04:21, Richard Newman wrote:
> I would suggest using sameAs or equivalentProperty, if the semantics
> are equivalent (as may well be the case for, e.g. messageIDs).
> Similarly, use subPropertyOf if that's the actual case (as you suggest
> for DC identifier).
<semanticHairSplitting>
I'd be practically careful, here... looking at all three:
* owl:equivalentClass [1]
* owl:equivalentProperty [2]
* owl:sameAs [3]
It looks like both equivalentClass and equivalentProperty actually may
not describe the relation you [Laurian] are after. Specifically, it
looks like both describe the state that the domain of the relations are
identical, but _specifically_not_ that the concepts are the same.
</semanticHairSplitting>
sameAs is defined for what you're after ["The owl:sameAs statements are
often used in defining mappings between ontologies"].
However, when you apply owl:sameAs to classes and properties, you've
started to treat classes as instances, which -- in my limited
understanding -- passes you from the computationally-feasible world of
OWL DL into the more heady world of OWL Full. I still need to hit the
DL books a couple more times to figure out exactly why.
Thus, I'd say "use owl:sameAs, and create your mapping for humans, not
automata".
rdfs:seeAlso is something else.
[1] http://www.w3.org/TR/owl-ref/#equivalentClass-def
[2] http://www.w3.org/TR/owl-ref/#equivalentProperty-def
[3] http://www.w3.org/TR/owl-ref/#sameAs-def
...jsled
--
http://www.asynchronous.org/ - `a=jsled; b=asynchronous.org; echo ${a}@${b}`
Received on Saturday, 10 July 2004 10:52:37 UTC