I well remember the joys of programming using the EBCDIC character set on IBM mainframes. That had the peculiarity that, if you printed the character set as a 16x16 grid, the alphabet was an 8x3 "block" in it. That meant that you couldn't test for an upper case letter using the normal "is it between A and Z" test because that range of character codes included other, non-alphabetic characters too
.
Ah, the "good old days"
