Zitat von Christophe Raffalli <Christophe.Raffalli@univ-savoie.fr>:

> Daniel de Rauglaudre a écrit :
> > Hi,
> >
> > On Tue, Aug 28, 2007 at 11:18:50PM +1000, Pietro Abate wrote:
> >
> >
> >> the other day I was looking for a fringe case to show me the worst
> >> complexity of the type inference algorithm...
> >>
> >
> > Interesting. To see the resulting time, I changed the code into:
> >
> >   let f =
> >   let x1 = fun x -> (x,x) in
> >   let x2 = fun y -> x1 ( x1 y ) in
> >   let x3 = fun y -> x2 ( x2 y ) in
> >   let x4 = fun y -> x3 ( x3 y ) in
> >   let x5 = fun y -> x4 ( x4 y ) in
> >   x5 ( fun z -> z ) ;;
> >
> >
> This is strange, it is problematic even with -rectypes while this is
> not necessary if the subtypes are shared ... I already so ocaml
> printing some types with sharing, so why is this not the case here ?
> (I am using 3.09.2)
[...]

Sorry, if I can't follow why this discussion was started.
Is this a purely-academic talk, or does it have any relevance
for programmers who want to solve a programming problem?

Where could the problem behind this kind of functions
occure in practical programming?

I tested the above code and it needed a lot of time
to execute in the toplevel. I stopped this then, before the
toplevel were ready.

I never had such a behavior of so long needed time
for any kind of code.
So I doubt that this is a problem that matters in the realworld.
But possibly I'm wrong. Can you please elaobrate on this?
Is this more than a academic mental exercise?

Ciao,
   Oliver