Question about Function Efficiency

[Matthew Donadio]

Hi all,

I want to create a function that creates a cascade of function
applications:

	cascade [ f1, f2, f3 ] x == (f3 . f2 . f1) x

It's a bit trivial, but it makes implementing flowgraphs (which are
usually drawn left-to-right) a bit easier.

I plan on profiling this, but which of the following do you think is the
most efficient implementation:

> cascade1 :: Num a => [[a] -> [a]] -> [a] -> [a]
> cascade1 []     = \x -> x
> cascade1 (f:fs) = cascade1 fs . f

> cascade2 :: Num a => [[a] -> [a]] -> [a] -> [a]
> cascade2 cs = foldr (.) (\x -> x) (reverse cs)

> cascade3 :: Num a => [[a] -> [a]] -> [a] -> [a]
> cascade3 cs = foldl (.) (\x -> x) (reverse cs)

Thanks.

-- 
Matthew Donadio <m.p.donadio@ieee.org>