The Power of Prolog [video]

Re:The Power of Prolog [video] by gambler on Monday May 11, 2020 @11:19AM (#23143087)

I would be gladly writing a lot of stuff in a Prolog-like language if there was a dialect that separated rule construction from traversal manipulation. This is such an obvious thing. I should be able to construct a set of rules and then say "but this time, ignore this branch" (without changing the tree with cuts). Or something like "do depth-first for up to 1000 levels and give me the incomplete set of answers".

In SQL you don't change the database simply because you want to run a different query. You manipulate the query, the indices and the query hints.

[ Parent | on HN ]

Re:The Power of Prolog [video] by falsissime on Monday May 11, 2020 @12:57PM (#23144281)

In fact, you can do this with Prolog already, however, you have restrict yourself to the pure monotonic subset. Many parts of Prolog the do not fit can be replaced by purer counterparts. See library(si), library(reif), library(clpz) for such attempts.

[ Parent | on HN ]

Re:The Power of Prolog [video] by rurban on Monday May 11, 2020 @04:00PM (#23146320)

I'm quite happy with picat, which supports traditional if and loops. It's also fast and supports some solvers, recursive backtracking is optional.

Skipping an expr by cut is in every prolog, but loops not.

[ Parent | on HN ]

Re:The Power of Prolog [video] by grizzles on Monday May 11, 2020 @10:58PM (#23150134)

You could try datalog with your own parser. Personally I feel like their is a need for a new industry standard language that lets you encode assertions with either a boolean or a probability.

[ Parent | on HN ]