Skip to Content.
Sympa Menu

coq-club - Re: [Coq-Club] problem with dependant types

coq-club AT inria.fr

Subject: The Coq mailing list

List archive

Re: [Coq-Club] problem with dependant types


chronological Thread 
  • From: Cuihtlauac ALVARADO <cuihtlauac.alvarado AT francetelecom.com>
  • To: Venanzio Capretta <Venanzio.Capretta AT mathstat.uottawa.ca>
  • Cc: Conor McBride <conor AT cs.rhul.ac.uk>, Nicolas Magaud <nmagaud AT cse.unsw.edu.au>, casteran AT labri.fr, june andronick <jandronick AT axalto.com>, coq-club AT pauillac.inria.fr
  • Subject: Re: [Coq-Club] problem with dependant types
  • Date: Thu, 18 Nov 2004 17:09:08 +0100
  • List-archive: <http://pauillac.inria.fr/pipermail/coq-club/>
  • Organization: France Telecom - R&D - MAPS/AST

> > What's so bad about K anyway?
> >
> >Dependent pattern matching is currently difficult in Coq, but it doesn't
> >have to be that way. The question is what to do about it.
> >
> >On conna\^it la chanson...
> >
> I don't have anything in particular agains K, but in general I don't 
> like axioms. When you are forced to use an axiom it means that your type 
> theory is not doing the job. One should rather rething the system so 
> that the desired properties are "structural" rather that axiomatic.
> In the case under study, Pierre Casteran showed how to do things without 
> assuming axioms, if you can do without them you should.
> A part from this philosophical point, I have no objection: go ahead and 
> use K!

On the other hand you can prove K for (vect A) if equality over A is
decidable, which happens, sometimes :-)

-- 
Cuihtlauac ALVARADO - France Telecom - R&D - +33296053273
2, av Pierre Marzin - 22307 Lannion - France




Archive powered by MhonArc 2.6.16.

Top of Page