- (*
- *
- * Note that from this abstract interface, the terms (expressions)
- * in the proof are not accessible at all; they don't need to be --
- * so long as we have access to the equivalence relation upon
- * proof-conclusions. Moreover, hiding the expressions actually
- * makes the encoding in CiC work out easier for two reasons:
- *
- * 1. Because the denotation function is provided a proof rather
- * than a term, it is a total function (the denotation function is
- * often undefined for ill-typed terms).
- *
- * 2. We can define arr_composition of proofs without having to know how
- * to compose expressions. The latter task is left up to the client
- * function which extracts an expression from a completed proof.
- *
- * This also means that we don't need an explicit proof obligation for 4.1.2.
- *)