0 | ||| A telescope is the variable context in a dependently typed program.
1 | ||| Dependent types with free variables are really only well-defined in a telescope.
2 | |||
3 | ||| A segment of a telescope is a tuple of types in that
4 | ||| telescope. These are usually implicit in formal developments, but
5 | ||| seem to be crucial to get a compositional account. Here we use
6 | ||| them to get 'n-ary dependent function' (compare with `Data.Fun`
7 | ||| and `Data.Rel`).
8 | |||
9 | ||| I've learned this from Conor McBride on an SPLV'19 bus ride.
10 | ||| A literary reference would be welcome. This paper is a start:
11 | ||| Unifiers as Equivalences: Proof-Relevant Unification of Dependently Typed Data.
12 | ||| Jesper Cockx, Dominique Devriese, Frank Piessens, ICFP'16.
13 | ||| but they don't seem to have segments and their left action on contexts.