-
@Thalesdisciple Many formulas about intervals which fail for points should probably be rewritten to say "open interval".
-
@Thalesdisciple e.g. the Cantor set contains no open intervals. Topologically this is better anyway: it's a nowhere dense set.
-
@Thalesdisciple e.g. if f is constant on an open interval, then f'=0 on that interval.
-
@Thalesdisciple Actually that fact is false for the endpoints anyway: the derivative could be 0 or undefined.