@joomy But the idea of “intrinsic verification” is certainly in the Epigram papers, the Power of Pi, etc. or anything that uses Vec A n as opposed to lists with predicates about the length. I don’t remember if those papers use that terminology though
@joomy you could have an “intrinsic semantics” in reynolds’ sense without using dependent types. The way we were using the terminology for twelf is the same as how I use it for agda
@joomy @Enetsee Like if Haskell didn't allow data T where C : A -> B -> T then would a constructor C : (A,B) -> T count A and B as "immediate" positions? What about Pair A B -> T for a user-defined Pair type?
@joomy @Enetsee I'd call the second one a nested type, and the first one "the grammar for inputs to datatype constructors includes the following type constructors:..."
RT appreciated:
I am happy to announce the official V1 of the Human-Centered Programming Languages textbook, the first book to fully blend PL foundations and human-centered computing with up-to-date research. Open access at: https://t.co/q5qimPR3U9
Paper: https://t.co/saNdJoPNtY