Superscripts (was Re: Roman Numerals (was Re: Improper grounds for rejection of proposal N2677))

From: Guy Steele (Guy.Steele@sun.com)
Date: Wed Nov 02 2005 - 10:44:40 CST

  • Next message: Philippe Verdy: "Re: Superscripts (was Re: Roman Numerals (was Re: Improper grounds for rejection of proposal N2677))"

    On Nov 2, 2005, at 1:39 AM, Jukka K. Korpela wrote:

    > ... Incidentally, when exponentiation is to be expressed
    > compactly in plain text, then I think UPWARDS ARROW U+2191 would
    > be a better symbol than circumflex ^ (which was originally taken
    > into use in exponential expressions since it can be imagined to be
    > a simulation of an upwards arrow).

    Agreed. I note, however, that ASCII originally (1963) did have an
    "upwards arrow"
    character, and a left arrow also. I well remember using Teletype
    Model 33
    terminals with these characters. There were programming languages
    in the 1960s that used left arrow for assignment and/or up arrow for
    exponentiation. (Digital Equipment Corporation's FOCAL languages, for
    example, which ran on their PDP-8 computers, used = for assignment
    but used up arrow for exponentiation.) When ASCII was revised
    in 1967, the up arrow was replaced with the circumflex and the
    left arrow with an underscore. These at least had some vague visual
    similarity to the old characters. The same old software continued
    to be used, of course; I also well remember how offended I was
    aesthetically when forced to deal with _ as an assignment character
    ---my code looked terrible.

    So I submit that the modern use of circumflex for exponentiation
    probably does not stem directly from its visual appearance---rather,
    it stems from the legacy of the use of up arrow for exponentiation
    in the 1960s and the replacement of this up arrow with the circumflex
    in ASCII. That the circumflex has some vague visual similarity to an
    up arrow has made its continued use over the years for exponentiation
    somewhat palatable.

    --Guy Steele



    This archive was generated by hypermail 2.1.5 : Wed Nov 02 2005 - 10:46:45 CST