In lemma A.15 of the paper proof, case T-Let (t = let p=t1 in t2), subcase "t1 is a value", you write: "Then t ---> match(p,t1)t2 by E-LetV". Isn't this a bit fast? It does not seem obvious to me from the proof that match(p,t1)t2 is indeed defined. -- Jerome