(Sometimes Gmail takes out white space so I'm hoping this post is indented properly.) I think that "numeric_cast" should be able to cast to and from enum types, so that it can be used as follows: #include <cstdint> #include <iostream> using std::cout; using std::endl; enum class Toad : std::uint8_t { Purple, Orange = 0x7f }; enum class Frog { Blue, Yellow, Pink = 0xFFFFu }; int main(void) { try { numeric_cast<Toad>( Frog::Pink ); } catch(std::exception const &e) { cout << e.what() << endl; } std::uint32_t my_32_bit_var = 0xffffffffU; try { numeric_caststd::uint16_t(my_32_bit_var); } catch(std::exception const &e) { cout << e.what() << endl; } try { numeric_caststd::uint8_t( Frog::Pink ); } catch(std::exception const &e) { cout << e.what() << endl; } try { numeric_cast<Toad>(0x1FFu); } catch(std::exception const &e) { cout << e.what() << endl; } } In order to achieve this behaviour, I've created my own "numeric_cast" in the global namespace as follows: #include // is_enum, underlying_type #include // numeric_cast template struct numeric_cast_helper { typedef T type; }; template<typename T> struct numeric_cast_helper { typedef typename std::underlying_type<T>::type type; }; template A numeric_cast(B const b) // This function takes its argument by value, and also returns by value -- which is fine for arithmetic types { return static_cast<A>( boost::numeric_cast::type> ( static_cast::type>(b) ) ); }