- From: Simon Spero <sesuncedu@gmail.com>
- Date: Fri, 26 Dec 2014 09:52:17 -0500
- To: Ignazio Palmisano <ipalmisano.mailings@gmail.com>
- Cc: public-owl-dev@w3.org
- Message-ID: <CADE8KM6MuQjQ8AnNtSdWYtMO7Vc0NwX+0Ry+gtMfb50tdOsjFw@mail.gmail.com>
It's not even clear that this is an anti-pattern.
There are cases where having the effective range of a property for a given
class be Nothing is exactly what you want.
An obvious case would be when instances of a subclass cannot have a part
that is permitted, but not required, for instances of their superclass.
This could be handled using cardinality constraints, but an explicit range
of Nothing is also viable.
For example: cat can has a tail. Manx cats don't have tails.
Cat subClassOf hasTail max 1
ManxCat subClassOf Cat
ManxCat subclassof hasTail max 0
or
ManxCat subClassOf hasTail only Nothing
If a class has multiple superclasses, it may be a combination of
restrictions that rules out the presence of any values for a property.
Suppose that all gluten-free biscuits contain nuts, and all nut-free
biscuits contain gluten (ie the two classes are disjoint).
A gluten-allergic person can eat only gluten-free biscuits.
A nut-allergic person can eat only nut-free biscuits.
A person who is both gluten-allergic and nut-allergic can eat only biscuits
which are both gluten-free and nut-free. Since the classes are disjoint,
there are no such things, and the poor unfortunate is doomed to go
biscuitless.
Being unable to eat biscuits is a sad form of existence, but nonetheless
existence it is.
To condemn these sufferers further to the pain of contradiction is as
unethical as it is illogical.
Simon
---
On Dec 25, 2014 12:53 PM, "Ignazio Palmisano" <ipalmisano.mailings@gmail.com>
wrote:
>
> On 25 December 2014 at 17:57, Leila Bayoudhi <bayoudhileila@yahoo.fr>
wrote:
> > I adopted the idea from this link. He and other works consider this case
> > leading to unsatisfiable classes.
> > this is the idea of this anti-pattern:
> > "The ontology developer created a universal restriction to say that C1
> > instances can only be linked with property R to C2 instances. Next, a
new
> > universal restriction is added saying that C1 instances can only be
linked
> > with R to C3 instances, with C2 and C3 disjoint. In general, this is
because
> > the ontology developer forgot the previous axiom in the same class or
in the
> > parent class."
> >
>
> This is an antipattern (i.e., an error) under the assumption that the
> property R is intended to be used for an individual - which is the
> common case. But, as Bijan remarked earlier, if R is not used in any
> assertion for an individual belonging to this restriction, then
> there's nothing wrong with the individual.
> Such a construct is /likely/ to be an error (it's a very complicated
> way to say R cannot be used for an individual), but it might be
> intended behaviour, and it's not forbidden by any rules.
> Cheers,
> I.
>
> >
> > Le Jeudi 25 décembre 2014 17h22, Uli Sattler
> > <Ulrike.Sattler@manchester.ac.uk> a écrit :
> >
> >
> > ...from what i can see you, you only have 2 "allValuesFrom"
restrictions on
> > Person - so it can't possibly be inconsistent (as long as I don't have
any
> > hamsters, it is perfectly uncontradictory to state that "all my
hamsters are
> > black" and "all my hamsters are not black...this will only cause an
> > inconsistency if I would also say "I have some hamsters", which would
then
> > need to be both black and not black).
> >
> > So, I repeat my earlier suggestion to read up more on owl semantics...
> >
> > Cheers, Uli
> >
> > On 25 Dec 2014, at 15:48, "Leila Bayoudhi" <bayoudhileila@yahoo.fr>
wrote:
> >
> > What I model it isn't equivalent to the OWL antipattern
> > OnlinessIsLoneliness(OIL)?
> >
> >
> > Le Jeudi 25 décembre 2014 16h34, Uli Sattler
> > <Ulrike.Sattler@manchester.ac.uk> a écrit :
> >
> >
> > Hi Leila
> >
> > this ontology is consistent - and even the class person is satisfiable:
it
> > only becomes unsatisfiable once you request that a person must have an
> > Id...but you don't seem to say that yet,
> >
> > Since this mailing list is read by humans, it would be good to post your
> > examples in Manchester syntax - the syntax you use is designed for
> > computers...
> >
> > Also, I strongly suggest you look more closely into Owl's model based
> > semantics - to get a better understanding of it. All the best!
> >
> > Cheers, Uli
> >
> > On 25 Dec 2014, at 11:38, "Leila Bayoudhi" <bayoudhileila@yahoo.fr>
wrote:
> >
> > Hi,
> > I intentionally want to raise inconsistency in my ontology by creating
these
> > axioms
> > <owl:Class
> > rdf:about="
http://www.semanticweb.org/dell/ontologies/2014/11/untitled-ontology-121#person
">
> > <rdfs:subClassOf>
> > <owl:Restriction>
> > <owl:onProperty
> > rdf:resource="
http://www.semanticweb.org/dell/ontologies/2014/11/untitled-ontology-121#hasIdentifier
"/>
> > <owl:allValuesFrom rdf:resource="&owl;rational"/>
> > </owl:Restriction>
> > </rdfs:subClassOf>
> > <rdfs:subClassOf>
> > <owl:Restriction
> > <owl:onProperty
> > rdf:resource="
http://www.semanticweb.org/dell/ontologies/2014/11/untitled-ontology-121#hasIdentifier
"/>
> > <owl:allValuesFrom rdf:resource="&xsd;dateTime"/>
> > </owl:Restriction>
> > </rdfs:subClassOf>
> > </owl:Class>
> > However, Hermit says that all is fine: the ontology is consistent(
even the
> > value spaces of the two data ranges are disjoint)
> > Can you tell me please why?
> >
> >
> >
> >
> >
>
On 25 December 2014 at 17:57, Leila Bayoudhi <bayoudhileila@yahoo.fr>
wrote:
> I adopted the idea from this link. He and other works consider this case
> leading to unsatisfiable classes.
> this is the idea of this anti-pattern:
> "The ontology developer created a universal restriction to say that C1
> instances can only be linked with property R to C2 instances. Next, a new
> universal restriction is added saying that C1 instances can only be linked
> with R to C3 instances, with C2 and C3 disjoint. In general, this is
because
> the ontology developer forgot the previous axiom in the same class or in
the
> parent class."
>
This is an antipattern (i.e., an error) under the assumption that the
property R is intended to be used for an individual - which is the
common case. But, as Bijan remarked earlier, if R is not used in any
assertion for an individual belonging to this restriction, then
there's nothing wrong with the individual.
Such a construct is /likely/ to be an error (it's a very complicated
way to say R cannot be used for an individual), but it might be
intended behaviour, and it's not forbidden by any rules.
Cheers,
I.
>
> Le Jeudi 25 décembre 2014 17h22, Uli Sattler
> <Ulrike.Sattler@manchester.ac.uk> a écrit :
>
>
> ...from what i can see you, you only have 2 "allValuesFrom" restrictions
on
> Person - so it can't possibly be inconsistent (as long as I don't have any
> hamsters, it is perfectly uncontradictory to state that "all my hamsters
are
> black" and "all my hamsters are not black...this will only cause an
> inconsistency if I would also say "I have some hamsters", which would then
> need to be both black and not black).
>
> So, I repeat my earlier suggestion to read up more on owl semantics...
>
> Cheers, Uli
>
> On 25 Dec 2014, at 15:48, "Leila Bayoudhi" <bayoudhileila@yahoo.fr> wrote:
>
> What I model it isn't equivalent to the OWL antipattern
> OnlinessIsLoneliness(OIL)?
>
>
> Le Jeudi 25 décembre 2014 16h34, Uli Sattler
> <Ulrike.Sattler@manchester.ac.uk> a écrit :
>
>
> Hi Leila
>
> this ontology is consistent - and even the class person is satisfiable: it
> only becomes unsatisfiable once you request that a person must have an
> Id...but you don't seem to say that yet,
>
> Since this mailing list is read by humans, it would be good to post your
> examples in Manchester syntax - the syntax you use is designed for
> computers...
>
> Also, I strongly suggest you look more closely into Owl's model based
> semantics - to get a better understanding of it. All the best!
>
> Cheers, Uli
>
> On 25 Dec 2014, at 11:38, "Leila Bayoudhi" <bayoudhileila@yahoo.fr> wrote:
>
> Hi,
> I intentionally want to raise inconsistency in my ontology by creating
these
> axioms
> <owl:Class
> rdf:about="
http://www.semanticweb.org/dell/ontologies/2014/11/untitled-ontology-121#person
">
> <rdfs:subClassOf>
> <owl:Restriction>
> <owl:onProperty
> rdf:resource="
http://www.semanticweb.org/dell/ontologies/2014/11/untitled-ontology-121#hasIdentifier
"/>
> <owl:allValuesFrom rdf:resource="&owl;rational"/>
> </owl:Restriction>
> </rdfs:subClassOf>
> <rdfs:subClassOf>
> <owl:Restriction
> <owl:onProperty
> rdf:resource="
http://www.semanticweb.org/dell/ontologies/2014/11/untitled-ontology-121#hasIdentifier
"/>
> <owl:allValuesFrom rdf:resource="&xsd;dateTime"/>
> </owl:Restriction>
> </rdfs:subClassOf>
> </owl:Class>
> However, Hermit says that all is fine: the ontology is consistent( even
the
> value spaces of the two data ranges are disjoint)
> Can you tell me please why?
>
>
>
>
>
Received on Friday, 26 December 2014 14:52:45 UTC