> (snip)

> > proper_divisors :: Int -> [Int]
> > proper_divisors n = [d | d <- [1 .. (floor . sqrt) n], mod n d == 0]
> (snip)
> >     No instance for (RealFrac Int)
> >       arising from a use of `floor' at pd.hs:2:37-41
> (snip)
> >     No instance for (Floating Int)
> >       arising from a use of `sqrt' at pd.hs:2:45-48

> the (floor . sqrt) can't accept an n :: Int, because they need n's type
> to be instances of Floating and RealFrac, and Int isn't.

> Prelude> :type sqrt
> sqrt :: (Floating a) => a -> a

> ... note that the return type is the same as the argument type. From a
> commonsense point of view, one can see part of the reason for the
> complaint about Int as being that for things like sqrt (2::Int) you
> couldn't return an Int answer (at least a correct one!).

> Of course,

> Prelude> :type mod
> mod :: (Integral a) => a -> a -> a

> mod expects an integral n, and you already declared n to be Int.

> How about,

> proper_divisors n = [d | d <- [1 .. (floor . sqrt . fromIntegral) n], mod n d == 0]

> By the way, though I actually much prefer your convention, upper and
> lower camel case seem to have established themselves as the Haskell way,
> so normally one would name your function properDivisors.

> Mark