>Bitwise operators are defined but don't seem to be documented. A rationale for their inclusion would be nice too. These look surprising to me - you can't use bitwise operators on binary floating point, so why here?