On Mon, 2004-12-13 at 22:38, Martin Berger wrote:
> it is quite simple: of course the monadic IO that has been discussed
> has side-effects. hence it is not funcitonal. but -- and that's a big
> but -- what the monadic types do is: LIMIT composability in a way that
> no typable context can OBSERVE the statefulness of the monadic code.
> an external observer, not bound to the typing discipline, can observe
> the stateful behaviour.

AH! Thank you. No *typable* context can *observe*...

Now we're getting closer to a more general characterisation...

> more generally: the question of whether a code fragment is  functional
> or not is meaningless without specifying the class of admissable
> observations,

Yay! Yes! Spot on! <lol>


-- 
John Skaller, mailto:skaller@users.sf.net
voice: 061-2-9660-0850, 
snail: PO BOX 401 Glebe NSW 2037 Australia
Checkout the Felix programming language http://felix.sf.net