Hmm, I guess you're right. Also EBCDIC does have ^ apparently, though not ↑: https://en.wikipedia.org/wiki/EBCDIC#Code_page_layout https://en.wikipedia.org/wik

Hmm, I guess you're right. Also EBCDIC does have ^ apparently, though not ↑: https://en.wikipedia.org/wiki/EBCDIC#Code_page_layout https://en.wikipedia.org/wiki/EBCDIC#Code_page_layout But IBM's BCD character sets, including the 48-character ones you allude to, didn't: https://en.wikipedia.org/wiki/BCD_(character_encoding)#Examples_of_BCD_codes https://en.wikipedia.org/wiki/BCD_(character_encoding)#Examp... (though Honeywell's did) There are a lot of decisions in Fortran that stem from the absence of useful characters. .LT., .LE., .EQ., .NE., .GT., and .GE. is another.