RFC: general sequences

[Ross Paterson]

On Mon, May 23, 2005 at 07:24:50PM +0200, Benjamin Franksen wrote:
> On Monday 23 May 2005 17:35, Ross Paterson wrote:
> >
> > 	foldr_l :: (a -> b -> b) -> b -> Seq a -> b
> > 	foldr_l f z xs = case viewL xs of
> > 		EmptyL -> z
> > 		x :< xs' -> x `f` foldr_l f z xs'
> >
> > The current definition is equivalent to [this] (but operates
> > directly on the internal structure).
> 
> I wonder: Can a compiler optimize the above so that it is as efficient 
> as the 'real' version? If not, why? If yes, everybody could easily 
> create their own folds.

No, they're too different in structure for a compiler to work out the
transformation.  People can still create their own folds -- they just
lose a constant factor over the insiders.