Re: Fwd: [Caml-list] Polymorphic Variants

[Tom <tom.primozic@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 2497 bytes --]

On 19/01/07, Dirk Thierbach <dthierbach@gmx.de> wrote:
>
> [Please don't reply directly to me AND to the list, just to the list
> is enough. Thanks.]


Sorry... didn't mean to... I mean, I didn't know there's a problem. I don't
get mail that's sent to me AND the caml-list twice...

On Fri, Jan 19, 2007 at 11:35:36AM +0100, Tom wrote:
> > On 19/01/07, Dirk Thierbach <dthierbach@gmx.de> wrote:
>
> > Well, in some sense, generic value overloading is somewhat like
> Haskell's
> > type classes,
>
> Yes. That's exactly what type classes are for -- overloading of functions
> (including infix operators).


No, no, you misunderstood me. Type-classes are run-time dispatched. What I
want is static/compile-time overloading. I meant they are similar for the
type inference system. My type system should use roughly the same algorithm
for inference that Haskell uses, but it should infer the typeclasses (the
possible types a type variable can be instantiated) automatically.

Actually, I don't even want such types be extensible - one function will
always have either ONE input type (int -> int), or ANY input types ('a ->
'a), but two such functions might have the same NAME. The compiler's task
isn't to supply all possible implementations of some mathematical function -
that's the programmers task:

if I define
  let negate a = - a
it will not have all possible types (int, float, complex, ...) but only one
of them. For another type, I have to define it again:
  let negate a = - a
I can also specify types:
  let negate a : complex = - a
(and of course ask the compiler to do it for me:
  overload let negate a = - a
but that's only a shortcut - internally, it would be expanded into n
functions, each ranging over one of n types. )

My overloading would be more C++ / C# / Java-like - but the inference
technique should be similar to Haskell's.

And that's exactly what the Haskell compiler does (see below).




BTW,
> the important thing is the type variable after the predicate, just
> ignore the forall's for the moment. So the above type is nonsense, a
> possible example type could look like:


Sorry, I don't know Haskell that well. But you know what I meant :)


> > Can Haskell overload values? And functions by their return type?
>
> It can overload integer and fixed point literals (in the way described
> above). It cannot overload any other "values" (whatever that should
> be :-).


By overloading values, I mean:

let null = 0
let null = false

if null then 1 else null

- Tom

[-- Attachment #2: Type: text/html, Size: 3891 bytes --]