coq-club AT inria.fr
Subject: The Coq mailing list
List archive
- From: Jean-Christophe Filliatre <filliatr AT lri.fr>
- To: wamelen AT marais.math.lsu.edu
- Cc: coq-club AT pauillac.inria.fr
- Subject: Re: Question about types
- Date: Wed, 29 Jul 1998 16:40:17 +0200 (MET DST)
> Variable U:Type
> Variable A:(Ensemble U)
>
> Is there a way to define a new type such that only the elements of A
> has that type?
>
> At first I thought that
>
> Record A_T : Type :=
> A_t {
> A_element : U;
> A_definition : (In U A A_element)}
>
> did it, but we want
>
> (a,b : A_T) a==b <-> (A_element a) == (A_element b)
>
> and I don't want to assume this as an Axiom.
The last property is not provable since there are several proofs of
(In U A A_element) in general, and consequently it is not true that
dependent tuples are equal if their first projections are.
But your problem is a recurrent one. You can find related questions
and answers in the archive of the Coq mailing list at
http://pauillac.inria.fr/coq/mailing-lists/coqclub/index.html
Among others, a particular discussion related to your problem is this one
initiated by Pierre Casteran:
http://pauillac.inria.fr/coq/mailing-lists/coqclub/0167.html
I think that the answer Christine Paulin gave will answer your
question. I hope so...
Best regards,
--
Jean-Christophe FILLIATRE
mailto:Jean-Christophe.Filliatre AT lri.fr
http://www.lri.fr/~filliatr
- Question about types, Paul van Wamelen
- Re: Question about types, Jean-Christophe Filliatre
Archive powered by MhonArc 2.6.16.