W3C home > Mailing lists > Public > www-webont-wg@w3.org > February 2003

Re: OWL syntax

From: Jeremy Carroll <jjc@hpl.hp.com>
Date: Tue, 25 Feb 2003 22:57:11 +0100
To: www-webont-wg@w3.org
Message-Id: <200302252257.11210.jjc@hpl.hp.com>


Jeremy:
>  _:a owl:equivalentClass _:b .
>  _:c owl:equivalentClass _:d .
> 
> (over four distinct bnodes, rather than three distinct
> bnodes in the original).

Peter:
>    The two versions entail each other, but are
> different syntactically.

which is the essence of my various suggested fixes -
force the concrete syntax only to talk about pairs
of descriptions
and these are semantically equivalent to abstract syntax 
n-tuples of descriptions

Possibilities include:
A: restrict abstract syntax to pairs in EquivalentClasses
and DisjointClasses
B: only allow for mapping pairs of EquivalentClasses or
DisjointClasses (n-tuples allowed in abstract syntax but
have no mapping under the mapping rules)
C: Use a variant of the n-tuple mapping rule that maps
an n-tuple into n or n(n-1)/2 pairs, and then map them
i.e.

EquivalentClasses(d1,d2,...dn)
=>
T(EquivalentClasses(di,d(i+1)))
   i=1,...n-1

I am sure there are others.
It is a bit clunky to not allow trees of EquivalentClasses
in the concrete syntax i.e. prohibit:

_:a owl:equivalentClass _:b .
_:a owl:equivalentClass _:c .

But I can't see how to do that well. 
(Note it is nice to get owl:disjointWith and owl:equivalentClass
behaving similarly - owl:disjointWith is distinctly harder).

Jeremy
Received on Tuesday, 25 February 2003 16:56:26 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:57 GMT