The uncertainty seems appropriate for the popular difficulties
controlling strictness: we're uncertain whether or not thunks have
been evaluated.
On the other hand, I'm a bit put off because Schrodinger's uncertainty
has strong overtones of nondeterminism, which would be very misleading
for people to assume about lazy evaluation.
Back on the first hand, It does seem very appropriate for unsafe IO!
Using unsafePerformIO creates a nondeterministic superstate that
collapses the first time you actually force the thunk-action.
Fun metaphor.
On 6/11/07, Andrew Coppin
"Data in Haskell is like Schrodinger's famous undead cat - it doesn't 'exist' until you 'obverse' it."
I just thought I'd share this useful (?) metaphore for describing to people what lazy evaluation is all about. ;-)
_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe