On Mon, 2007-07-16 at 02:29 +0100, Conor McBride wrote:
Hi
Whoops! Only just spotted this. Many apologies.
On 10 Jul 2007, at 20:35, Creighton Hogg wrote:
Me:
No, an operating system is supposed to remain responsive. And that's what total coprograms do.
I'm sorry, but can you expand a little further on this? I guess I don't understand how a corecursion => responsive to input but not terminating. Where does the idea of waiting for input fit into corecursion?
You'll be needing a bit of higher-order corecursion for that. Here's a coprogram for haskell-cafe:
data{-codata-} Punter = Speak String (String -> Punter)
A Punter is guaranteed to ask a question, and whatever answer you give them, they've always got another question, forever! Meanwhile, a String -> Punter is a good-natured soul, always up for answering questions. Hence haskell-cafe is a productive coprogram(*) producing a stream of questions and answers!
data{-codata-} Stream x = x :> (Stream x)
cafe :: Punter -> (String -> Punter) -> Stream (String, String) cafe (Speak question learn) guru = let Speak answer guru' = guru question in (question, answer) :> (cafe (learn answer) guru')
If the Punter asks the appropriate question, perhaps the guru will spend the rest of time thinking about an answer.