You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Right now, rify can't do IO. It can't do math on numeric literals. It can't answer questions like is 1 < 2.
Perhaps it could. It would need a way to message its caller, asking for the answers to specific questions. Processing logic would become a conversation between the caller, and the reasoner. This is similar to chalk's execution model.
The text was updated successfully, but these errors were encountered:
bddap
changed the title
Generic logical predicates.
Generic predicates.
Dec 1, 2020
When the reasoner is "Curious" about something. It could request information from the caller by deriving a fact.
if [?a rdf:type Did] and [?a mayClaim ?b] then [?a type PleaseDeref]
When the caller sees a new triple of the form [?a type PleaseDeref], it would start a did lookup for ?a. Reasoning would continue while ?a is being dereferenced. After the dereferenced value of '?a' is known, it would be converted to something like explicit ethos form, then fed back into the reasoner to "Satisfy its Curiosity".
The proof verification API is generic enough that it should be possible to include generic predicates as premises to rule applications. The verifier would need to check generic predicates when they aren't present in the already established claimgraph.
Right now, rify can't do IO. It can't do math on numeric literals. It can't answer questions like
is 1 < 2
.Perhaps it could. It would need a way to message its caller, asking for the answers to specific questions. Processing logic would become a conversation between the caller, and the reasoner. This is similar to chalk's execution model.
The text was updated successfully, but these errors were encountered: