Re: owl:Thing (was Re: Managing Co-reference)

2008/5/21 Jeremy Carroll <jjc@hpl.hp.com>:
>
> Thomas Bandholtz wrote:
>>
>> suddenly so quiet here :-)
>>
>> Re-reading the OWL spec I found that owl:sameAs has domain & range =
>> owl:Thing.
>> So any two concepts which are not derived from owl:Thing cannot be
>> connected by owl:sameAs.
>
>
> My view is that it was an error for the terms  owl:Thing and owl:Class to
> have ever been introduced. Instead I believe it would have been less
> confusing to use rdfs:Resource and rdfs:Class even in OWL DL.
>
> In OWL Full, the two concepts owl:Thing and owl:Class are just a confusion;
> they are equivalent classes to rdfs:Resource and rdfs:Class.
>
> In OWL DL - the syntactic restriction mean that it is impossible to express
> the fact that the concepts owl:Thing and owl:Class are narrower than
> rdfs:Resource and rdfs:Class; but stepping outside OWL DL, to a very human
> but formally non-existent meta-level, it is clear that they are.
>
> If OWL DL had used the rdfs terms, then the difference could have been
> articulated as, well of course, owl:Nothing is an rdfs:Resource, but you
> can't say that in DL.
>
> Jeremy

Defining a thing as everything which is nothing as well as everything
which is not-nothing doesn't exactly give you many semantic ideas,
considering that Nothing derives from Thing as everything which is a
thing which could be nothing or not-nothing. How can a useful
vocabulary start from this circular, spiralling, reference. "There was
a Thing, and together with all of Nothing and Not-Nothing. And there
was Nothing, which was everything which was not of a Thing, which was
all of Nothing and Not-Nothing" It just doesn't seem to fit that
Nothing is explicitly included in the description of Thing given that
it is elsewhere described as a direct complement. Is Nothing ever used
in OWL for a practical purpose or is Thing the preferred class to use.
If Nothing is "the empty Class" then it can never be derived from or
be added to, because only Thing's are instances of actual classes with
corresponding properties if I am not mistaken.

[http://www.w3.org/2002/07/owl]

<Class rdf:ID="Thing">
  <rdfs:label>Thing</rdfs:label>
  <unionOf rdf:parseType="Collection">
    <Class rdf:about="#Nothing"/>
    <Class>
      <complementOf rdf:resource="#Nothing"/>
    </Class>
  </unionOf>
</Class>

<Class rdf:ID="Nothing">
  <rdfs:label>Nothing</rdfs:label>
  <complementOf rdf:resource="#Thing"/>
</Class>

Sorry if this sounds like uninformed rambling and I am missing an
important point why these two Classes need to obliquely reference each
other using complement in this way as the basis for a consistent
ontological system.

Peter

Received on Wednesday, 21 May 2008 11:58:57 UTC