-
Notifications
You must be signed in to change notification settings - Fork 9
SLD Resolution? #209
Comments
Hi @brurucy, We used SLD resolution to implement Prolog-style extensions. For example, Modus allows writing rules like
where Thank you! Yes, we are continuing working on Modus, but there have not been any concrete publications/releases yet. What is your opinion about this direction? Do you have any ideas or specific usecases? |
I think this could be solved in different ways.
To my understanding, this means that by the time the head is evaluated, all "substitutions" that are applied to it will indeed yield a ground fact, hence not needing support for existential variables at all. I think "prolog-style" would be a "more correct" label if that variable could be propagated further than the current rule. Why couldn't the rule be:
I suppose that a "user query" could be modelled as an insertion to that "cflags" predicate. It could also be done through skolemization, where
We could chat more about this. My academic email is: [email protected] |
Hi,
What's the reason that lead you to use a prolog evaluation strategy (SLD resolution) for a datalog?
For instance,
magic-sets
is a relatively simple datalog-first strategy that would most certainly outperform SLD resolution by orders of magnitude.The project is awesome though! Has there been a follow-up?
The text was updated successfully, but these errors were encountered: