- From: Ignazio Palmisano <ipalmisano.mailings@gmail.com>
- Date: Tue, 10 Mar 2015 19:23:55 +0000
- To: "public-owl-dev@w3.org" <public-owl-dev@w3.org>
On 10 March 2015 at 18:56, Bijan Parsia <bijan.parsia@manchester.ac.uk> wrote: > On 10 Mar 2015, at 14:08, Leila Bayoudhi <bayoudhileila@yahoo.fr> wrote: > > Hi, > Thanks all for your answers. > The proposed solution is taken from ontology design patterns site . > > > I cannot say that I find the description there particulalyr perspicuous. > > As the attached figure shows: > subClassOf( Carnivorous_plant unionOf(animal, plant)) and we already have > subClassOf(Carnivorous_plant , plant) > > > Ok. I don’t think this is what the pattern meant, as that is silly (because > redundant). > > I think that this is the transform: > Original ontology: > Animal ⊑Fauna-Flora, > Plant ⊑Fauna-Flora, > Carnivorous-Plant ⊑Plant, > Plant ⊑ Not(Animal) > new ontology: > Animal ⊑Fauna-Flora, > Plant ⊑Fauna-Flora, > Carnivorous-Plant ⊑(Plant or Animal), > Plant ⊑ Not(Animal) > > The solution: > Animal ⊑Fauna-Flora, > Plant ⊑Fauna-Flora, > Carnivorous-Plant ⊑Plant, > Carnivorous-Plant ⊑(Plant or Animal), > Plant ⊑ Not(Animal) > is pointless as it’s equivalent to the original ontology. > > BTW, this is a terrible pattern for this case. What this tells us is that > (potentially) *some carnivorous plants are animals and thus not plants*. But > that’s surely not the intent! > > Frankly, I don’t know why anyone would want carnivorous plants to be animals > :), but presumably the idea is that they are animalesque plants, not that > there are some that are non-plants. > My two cents worth of a guess are on the idea that by 'plant' what is intended is 'something that does not eat animals', autotroph, 'not a hunter', and that 'animal' stands for 'heterotroph', 'a hunter', 'something that eats animals or plants'. A carnivorous plant has characteristics of a plant, in that it /is/ one, but it also has characteristics of 'animal'. I'm waving hands a lot here, between my shaky understanding of biology and the shakiness of the example. It's clearly not a good example for this pattern, as we're not managing to make heads or tails of the example, let alone the pattern. I. > Now perhaps somewhere you have that all carnivores are animals. That *would* > be a problem, but the solution is obvious *in that case* and is in fact this > pattern. > > Does this at end means that subClassOf(Carnivorous_plant plant) (this is > according to the reasoner since subClass(A B) subclass(A C) ==> subClass(A > intersectionOf(B C)). > So is this solution correct? or does it propose a redundant axiom? > Thanks for answering me. > > > > Le Mardi 10 mars 2015 9h25, Bijan Parsia <bijan.parsia@manchester.ac.uk> a > écrit : > > > On Mar 10, 2015, at 0:47, "Leila Bayoudhi" <bayoudhileila@yahoo.fr> wrote: > > Hi, > Having already: > subClassOf(person animal) > DisjointClasses(woman animal ) > DisjointClasses(man animal) > > We want to introduce subClassOf (person ObjectUnionOf(woman man) > This may introduce inconsistency. > > > No but it does introduce in satisfiability. > > So, we choose as a solution to introduce a subClassOf (person > ObjectUnionOf(animal ObjectUnionOf(woman man)). > > > But this is not helpful if you have the original axioms. Ie they are > equivalent. > > According to protegé, the ontology is no longer inconsistent. However, it > seems as if the ontologist wants at the end to say that: > subClassOf (person ObjectUnionOf(woman man): > > > If you preserve the original axioms, then there will still be no person who > is either a man or a woman. > > Is it correct what i am saying? > If it is not: is it problem of my proposed solution for maintaining > consistency? > Am I introducing redundant axioms(though OWL 2 DOES NOT care for this, I > care). > > > Without the union with animal, person is unsat. > > With the union with animal, the new axiom is redundant. > > Thx for answering me those questions? > > > > > <Djedidi_LOP2_WOP09.pdf> > >
Received on Tuesday, 10 March 2015 19:24:24 UTC