Hi Everyone,

I'm not sure to get the idea behind the "impossible" keyword.
Is it an optional keyword (only for readability purposes) or are there any cases where it matters to the coverage checker (in terms of either performance or correctness)?

Cheers,

- Another related question is about the following example In page 278

twoPlusTwoNotFive : 2 + 2 = 5 -> Void
twoPlusTwoNotFive Refl impossible

I wonder if that means that if we replace 2 + 2 = 5 by any false hypothesis (over Nat) then the compiler will be able to decide in a general way if it's indeed false by typechecking (or refusing to) the following
FalseHypothesis NatEqProposition -> Void
FalseHypothesis Refl impossible

and if it generalizes to equality propositions over other data types (involving total functions)?