N+1 is the width of
uintmax_t, assuming most common representations of signed integers.
2 Well, there is exactly one exception to that: on systems where
-INTMAX_MIN is not representable in
uintmax_t, this same value may cause problems.
INT_MAX can never be 10000, because is it at least 215-1.
CHAR_BIT isn't 8 only on special hardware such as DSP, and if is is 8, the typedef
uint8_t must exist.