Hmmm... this is making me think of infinite list comprehensions in
Haskell, which are evaluated to the degree required when required.
Maybe a peek at the Haskell web site might provide some sort of answer
on this ? (It's been a while; I know you can define a term that would
theoretically result in an infinite output, but because Haskell is
lazily evaluated it doesn't try working out the entire data set at
definition, and only works out as much as it needs to answer whatever
function you compute. I also seem to recall that older functional
languages fudged I/O by pretending it was an infinite list of inputs
and an infinite list of outputs, but memory gets hazy about that !)
Received on Mon May 22 2006 - 08:07:38 CDT
