Congratulations on your article; I am new to GRAKN.AI and my first approach is via reading your article, which it has been most helpful. However, I have a couple of simple queries:
- You wrote “Note that the predicates are reified here (…)” when you say here, what do you mean? that predicates are reified in GRAKN.AI in comparison to ASP?;
- I wonder why defined relations have arbitrary arity. So, following your example, one could also apply relation “played” (of arity 3 in the example) to 4 arguments?;
- Finally - and out of curiosity- when writing inference rules inspired in Prolog “head :- body” format, rhs (which I assume means right-hand side) is a rule set containing the head whereas lhs is a rule set containing the body. It seems to me counter-intuitive to swap them when proclaiming that the knowledge system is inspired in Prolog, but I presume there is some obvious reasoning behind that I am missing. Is it perhaps based on the turnstile format of inference rules “body |- head”?
That’s all. Once again, congratulations to you and your team at GRAKN.AI.