This is functionally identical to the OCaml polymorphic inequality test <
except that it is total even on floating-point NaNs. More importantly, it will
more efficiently short-cut comparisons of large data structures where
subcomponents are identical (pointer equivalent).
FAILURE CONDITIONS
May fail when applied to functions.
EXAMPLE
# 1.0 < nan or nan < 1.0;;
val it : bool = false
# 1.0 nan;;
val it : bool = false
# nan 1.0;;
val it : bool = true