Under OS/390, floating-point representation (unlike scientific notation) uses a base of 16 rather than base 10. IBM mainframe systems all use the same floating-point representation, which is made up ...