Re: Pragmatic Problems in the RDF Ecosystem (Was: Re: Toward easier RDF: a proposal)

On Mon, Dec 10, 2018 at 11:25:48AM -0500, David Booth wrote:
> On 12/10/18 5:31 AM, Christopher Yocum wrote:
> > On Sun, 9 Dec 2018 23:56:29 -0500 David Booth wrote:
> > > What OWL2 features do you use, i.e., what inference constructs?
> > 
> > The easiest way to see that is to look at:
> > https://github.com/cyocum/irish-gen/blob/master/earlyIrishRelationship.ttl
> 
> That's very helpful.   One more question: I see that you use
> owl:differentFrom quite a lot.  Are you making use of that in a reasoner
> somewhere in your work?  In other words, would your application behave
> differently if you did not have the owl:differentFrom assertions in your
> RDF?  If so, how?

Not that I know of and I do not believe that it would behave
differently.  It was cargo cult code here because when I started I did
not quite understand everything so I just followed the example of
other ontologies which used it.  As it has continued to work and I
have not had any known issues with it, I have not prioritised changing
it.

> 
> BTW, I noticed one odd thing in your use of owl:inverseOf, that looks like
> it may be a mistake.   I see this:
> 
>   :friendOfGroup  owl:inverseOf :antagonistOfGroup .
> 
> Normally owl:inverseOf states that one predicate simply swaps the order of
> the subject and object of the inverse predicate, such as:
> 
>   f:hasHusband owl:inverseOf f:hasWife .
> 
> so that if you write:
> 
>   :fred f:hasWife :ginger .
> 
> you can infer:
> 
>   :ginger f:hasHusband :fred .
> 
> Similarly, if you had:
> 
>   :fred :friendOfGroup :ManchesterFans .
> 
> then a reasoner would infer:
> 
>   :ManchesterFans :antagonistOfGroup :fred .
> 
> and that doesn't seem like what you want.

You are correct and this is a bug that got past me during peer review.
I will need to go back and change this.  We need the friend/antagonist
relationships because they appear in our source material but it will
need to be modeled without the inverse as you have suggested.

Thank you for that!

All the best,
Chris

Received on Monday, 10 December 2018 20:15:04 UTC