Chris Hecker <checker@d6.com> a écrit :

> >I can't find a link to any papers by Okasaki on this subject.
> 
> http://citeseer.nj.nec.com/okasaki98view.html

One way to achieve okasaki goal is to provide in the module interface
a lazy concrete_type_of_abstract_type.

Thus, we could write Okasaki's example (page 15) this way:

module type SEQUENCE =
  sig
    type 'a t
    val empty : 'a t
    val cons : 'a -> 'a t -> 'a t
    val append : 'a t -> 'a t -> 'a t
    val lazy_list_of_sequence : 'a t -> 'a Stream.t
  end
      
module Sequence : SEQUENCE =
  struct
    type 'a t = Empty | Cons of 'a * 'a t | Append of 'a t * 'a t
	
    let empty = Empty
    let cons hd tl = Cons hd tl
    let append l1 l2 = Append l1 l2
	
    let rec lazy_list_of_sequence = function
      | Empty -> [< >]
      | Cons a b -> [< 'a ; lazy_list_of_sequence b >]
      | Append a b -> [< lazy_list_of_sequence a; lazy_list_of_sequence b >]
  end
    
open Sequence

let length l =
  let rec length_aux = parser
    | [< 'hd ; tl >] -> 
	1 + length_aux tl
    | [< >] -> 0 in
  length_aux (lazy_list_of_sequence l)
    
let l = append (cons 1 (cons 2 (cons 3 empty))) (cons 4 empty) in
  Printf.printf "%d\n" (length l)

It's syntactically heavy (on an easy example...), it's not
intellectually satisfying, it wastes memory, etc. but it
works :-)

Maybe some camlp4 syntactic sugar could do the trick.

-- 
  Selon les logs, il y a des pertes de porteuses, mais surtout des
  requêtes de déconnexion qui viennent de votre machine.
  Il faut se rappeler que windows 95 est un systeme bio-dégradable.
  -+- Support technique HOL in: Guide du Cabaliste Usenet - CQFD ! -+-
-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners