caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
* Re: Type issue
@ 2007-11-24  5:47 oleg
  0 siblings, 0 replies; 2+ messages in thread
From: oleg @ 2007-11-24  5:47 UTC (permalink / raw)
  To: jtbryant, Andrej.Bauer; +Cc: caml-list


As Alain Frisch mentioned, with polymorphic recursion, inference is in
general impossible. Luckily, polymorphic recursion can be integrated
into a HM system if we require a type annotation on polymorphically
recursive functions. The annotation can come in various forms: for
example, if we use recursive modules, the annotation is given in the
signature. It seems using first-class polymorphism of OCaml records
might be a bit simpler:

type 'a t = Cond of bool t * 'a t * 'a t | Value of 'a;;
  type 'a t = Cond of bool t * 'a t * 'a t | Value of 'a

type ft = {ft : 'a . 'a t -> 'a};;
  type ft = { ft : 'a. 'a t -> 'a; }

let f = 
  let rec f' = {ft = function 
    | Cond (c,t,e) -> if f'.ft c then f'.ft t else f'.ft e
    | Value x -> x}
  in f'.ft;;

 val f : 'a t -> 'a = <fun>

f (Cond ((Value true),(Value 1),(Value 2)));;
- : int = 1

One should think of 'ft' as marking the places where the big-lambda is
introduced and eliminated.  Luckily, the extended value restriction
(or should I say, restricted value restriction) helps us avoid
eta-expansion in the definition of f.


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Type issue
@ 2007-11-23  4:01 Jonathan T Bryant
  0 siblings, 0 replies; 2+ messages in thread
From: Jonathan T Bryant @ 2007-11-23  4:01 UTC (permalink / raw)
  To: caml-list

List,

I don't understand the following typing:

# type 'a t = Cond of bool t * 'a t * 'a t | Value of 'a;;
type 'a t = Cond of bool t * 'a t * 'a t | Value of 'a

# let rec f t = match t with
      Cond (c,t,e) -> if f c then f t else f e
    | Value x -> x
  ;;
val f : bool t -> bool = <fun>

I don't understand why f isn't of type 'a t -> 'a.  I understand that f 
is applied to a bool t, but I would think that that would just be a 
particular instantiation of 'a, not enough to fully determine its type.

--Jonathan Bryant


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2007-11-24  5:50 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2007-11-24  5:47 Type issue oleg
  -- strict thread matches above, loose matches on Subject: below --
2007-11-23  4:01 Jonathan T Bryant

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).