coq-club AT inria.fr
Subject: The Coq mailing list
List archive
- From: Roland Zumkeller <Roland.Zumkeller AT polytechnique.fr>
- To: coq-club AT pauillac.inria.fr
- Subject: [Coq-Club] Module problem
- Date: Thu, 13 May 2004 18:45:06 +0200
- List-archive: <http://pauillac.inria.fr/pipermail/coq-club/>
This is a simplified example that is only meant to demonstrate what I'd like to achieve.
Module Type Variables.
Variable V : Set.
End Variables.
An interpretation associates a natural number to every variable.
Module Type Interpretation.
Declare Module Vars : Variables.
Variable interp : Vars.V -> nat.
End Interpretation.
The extension of an interpretation allows also to interprete sums and products:
Module Interpretation_ext (I:Interpretation).
Inductive X : Set :=
| Plus : I.Vars.V -> I.Vars.V -> X
| Times : I.Vars.V -> I.Vars.V -> X.
Definition interp_ext (x : X) :=
match x with
| Plus v w => I.interp v + I.interp w
| Times v w => I.interp v * I.interp w
end.
End Interpretation_ext.
Now, I'd like to compare two different interpretations that have the same set of variables.
Module Interp_compat (V:Variables) (I1 I2 : Interpretation with Module Vars := V).
Module I1_ext := Interpretation_ext(I1).
Module I2_ext := Interpretation_ext(I2).
Definition same := forall x, I1_ext.interp_ext x = I2_ext.interp_ext x.
The last definition gets rejected:
Error: In environment
x : I1_ext.X
The term "x" has type "I1_ext.X" while it is expected to have type "I2_ext.X"
The type X in Interpretation_ext is the same for I1_ext and I2_ext
I also tried to specify this as a constraint when defining I1 and I2 :
Module Type IntExt.
Variable X : Set.
Variable interp_ext : X -> nat.
End IntExt.
Module Interp_compat (V:Variables) (I1 I2 : Interpretation with Module Vars := V).
Module I1_ext : IntExt := Interpretation_ext(I1).
Module I2_ext : IntExt with Definition X := I1_ext.X := Interpretation_ext(I2).
But I get the error :
"User error: Signature components for label X do not match"
Is there a way to do this ? Of course I1.V.Vars and I2.V.Vars are the same, but can this be extended to the results of functor applications ?
thank you
- [Coq-Club] Module problem, Roland Zumkeller
- Re: [Coq-Club] Module problem, Roland Zumkeller
Archive powered by MhonArc 2.6.16.