Martijn van Steenbergen <mart...@van.steenbergen.nl> writes:
>> inTwain as = let (x,y,_) = foldr (\a (r,s,t) -> case (t) of
>> {b:(b':bs) -> (r,a:s,bs); _ -> (a:r,s,t)}) ([],[],as) as in (x,y)
> This one is very interesting.
Yes, neat.
> I'm not too happy with the whole list as part of the initial
> state. That feels like cheating to me--although I obviously failed to
> specify this in my original question.
Can you avoid it? If you only allow one traversal of the input (and
I think the above might qualify as two, albeit simultaneous), how do
you know when you're halfway through?
I seem to remember one example case that regular languages can't solve
is a^nb^n -- is this for the same (or a related) reason?
-k
--
If I haven't seen further, it is by standing in the footprints of giants
_______________________________________________
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe
