The CPUs in the zSeries mainframe actually support 3 floating point formats (in fact there is like 10 of them if you count all the sizes): - IBM legacy hexadec

The CPUs in the zSeries mainframe actually support 3 floating point formats (in fact there is like 10 of them if you count all the sizes): - IBM legacy hexadecimal floating point you mentioned - IEEE 754 binary floating point - IEEE 754 decimal floating point (they were a first implementation of it) So perhaps you were downvoted because on the mainframe, you don't have to use the legacy format at all.