coq-club AT inria.fr
Subject: The Coq mailing list
List archive
- From: Alexandre Pilkiewicz <alexandre.pilkiewicz AT polytechnique.org>
- To: Michael Shulman <mshulman AT ucsd.edu>
- Cc: Coq Club <coq-club AT inria.fr>
- Subject: Re: [Coq-Club] typeclasses and sections
- Date: Mon, 21 Nov 2011 22:32:38 +0100
Context allows you to do this:
<<<<
Class inhabited (A:Type):= {repr:A}.
Generalizable Variable A.
Section FOO.
Context `{inhabited A}.
Definition unsafe_head (l:list A) :=
match l with
| nil => repr
| cons x _ => x
end.
End FOO.
About unsafe_head.
(*
unsafe_head : forall A : Type, inhabited A -> list A -> A
Arguments A, H are implicit and maximally inserted
Argument scopes are [type_scope _ list_scope]
unsafe_head is transparent
Expands to: Constant Top.unsafe_head
*)
>>>
On Mon, Nov 21, 2011 at 9:38 PM, Michael Shulman
<mshulman AT ucsd.edu>
wrote:
> Is it possible to declare a hypothesis in a section which will be
> automatically *implicitly* generalized upon completion of the section?
> In particular, I would like to do this for an instance declaration, so
> that
>
> Section sect.
>
> Hypothesis A : Type.
> Hypothesis inst : type_class A.
>
> Theorem thm : B -> C.
> ...
>
> End sect.
>
> would be equivalent to
>
> Theorem `{inst : type_class A} : B -> C.
>
> I suppose that I can do this with
>
> Implicit Arguments thm [[A] [inst]].
>
> (Is that literally the same thing, or is it going to miss out on some
> typeclass magic?) In any case, that is tedious if the section
> contains many definitions and theorems.
>
> Mike
>
- [Coq-Club] typeclasses and sections, Michael Shulman
- Re: [Coq-Club] typeclasses and sections, Robbert Krebbers
- Re: [Coq-Club] typeclasses and sections, Alexandre Pilkiewicz
- Re: [Coq-Club] typeclasses and sections,
Michael Shulman
- Re: [Coq-Club] typeclasses and sections,
Tom Prince
- Re: [Coq-Club] typeclasses and sections,
Michael Shulman
- Re: [Coq-Club] typeclasses and sections, Tom Prince
- Re: [Coq-Club] typeclasses and sections,
Michael Shulman
- Re: [Coq-Club] typeclasses and sections,
Tom Prince
- Re: [Coq-Club] typeclasses and sections,
Michael Shulman
Archive powered by MhonArc 2.6.16.