Greetings,
I ran a reasoner (HermiT) over FOLIO and found a few errors of which you might want to be made aware. (3) is a pervasive issue.
(1) Perkins Coie is inferred to be a Lawyer. This, of course, is not true. Perkins Coie is a law firm. The reason for this inference is that Perkins Coie is asserted as folio:represented Microsoft, and the domain of folio:represented is Lawyer. This means that, for all x, if x folio:represented y, then x is a lawyer. The solution for this is to change the domain of folio:represented to Lawyer or Law Firm.
(2) Google, LLC is inferred to be both a Litigant and a Legal Services Provider. I don't think this is true, but I am not an expert in the legal field. Regardless, the reason for this will have consequences I'm sure you don't want. The reason is that Google, LLC is asserted to have rdf:type 'folio:appreadedBefore some Magistrate' and folio:appearedBefore has the domain intersection of Litigant and Legal Services Provider. This means that, for all x, if x folio:appearedBefore y, then x is both a Litigant and Legal Services Provider. Surely, there are some cases where this is false. The solution for this is to change the domain to Litigant or Legal Services Provider.
(3) (2) points to a more pervasive issue, which is that many FOLIO object properties have domains and/or ranges that are intersections that will lead to unwanted inferences. Folio:appearedBefore, for example, also has range the intersection of 'Forums and Venues' and 'Decision Maker for Legal Domain.' This means that, for all y, if x folio:appearedBefore y, then y is both a 'Forums and Venues' and 'Decision Maker for Legal Domain.' Clearly, this consequence is false in all cases, since there is nothing that can be both a 'Forums and Venues' and a 'Decision Maker for Legal Domain.' Generally, the solution for this is to change the intersections to disjunctions, but will have to be judged on a case by case basis.
All the best,
Alec Sculley
Greetings,
I ran a reasoner (HermiT) over FOLIO and found a few errors of which you might want to be made aware. (3) is a pervasive issue.
(1) Perkins Coie is inferred to be a Lawyer. This, of course, is not true. Perkins Coie is a law firm. The reason for this inference is that Perkins Coie is asserted as folio:represented Microsoft, and the domain of folio:represented is Lawyer. This means that, for all x, if x folio:represented y, then x is a lawyer. The solution for this is to change the domain of folio:represented to Lawyer or Law Firm.
(2) Google, LLC is inferred to be both a Litigant and a Legal Services Provider. I don't think this is true, but I am not an expert in the legal field. Regardless, the reason for this will have consequences I'm sure you don't want. The reason is that Google, LLC is asserted to have rdf:type 'folio:appreadedBefore some Magistrate' and folio:appearedBefore has the domain intersection of Litigant and Legal Services Provider. This means that, for all x, if x folio:appearedBefore y, then x is both a Litigant and Legal Services Provider. Surely, there are some cases where this is false. The solution for this is to change the domain to Litigant or Legal Services Provider.
(3) (2) points to a more pervasive issue, which is that many FOLIO object properties have domains and/or ranges that are intersections that will lead to unwanted inferences. Folio:appearedBefore, for example, also has range the intersection of 'Forums and Venues' and 'Decision Maker for Legal Domain.' This means that, for all y, if x folio:appearedBefore y, then y is both a 'Forums and Venues' and 'Decision Maker for Legal Domain.' Clearly, this consequence is false in all cases, since there is nothing that can be both a 'Forums and Venues' and a 'Decision Maker for Legal Domain.' Generally, the solution for this is to change the intersections to disjunctions, but will have to be judged on a case by case basis.
All the best,
Alec Sculley