Hi,
I stopped reading the list for a while ... Simon Peyton Jones kindly
forwarded the original message (and I re-subscribed to the list :-)
Anyway, last summer I've played a little with modeling QC in Haskell
using the high-level ideas you describe. I have nothing yet to show
but meanwhile you
Background. We have two notions of state (ST and IO) with two
mediating operations:
unsafeIOtoST :: IO a -> ST s a
stToIO :: ST s a -> IO a
unsafeIOtoST is documented as unsafe because exceptions that would
have been caught in the IO monad are not caught in the ST monad but
this is irrelevant to
