RE: Characters consisting of vertical lines; Possible attempts to encode tally marks

From: Philippe Verdy (
Date: Thu Mar 01 2007 - 06:36:49 CST

  • Next message: Guy Steele: "Re: Characters consisting of vertical lines; Possible attempts to encode tally marks"

    > De la part de Andrew West
    > The tally method of writing the strokes of the five-stroke character 正
    > zheng (U+6B63) is very widely used in China and Japan, but I don't
    > really see any necessity to encode a series of characters:
    > ZHENG STROKES 1 2 3 AND 4
    > But how about encoding an ENEMY KILL TALLY MARK:
    > <>
    > <>

    Really not needed. In fact you could count anything this way, and you could
    use any symbol for such tally... So there are:


    Dispose them in groups as much as can fit (your examples show grouping by
    rows of 6 or 7, but it could equivalently be grouped by rolumns). Really the
    character used does not matter, it's obvious that it is an association
    between a tally count, and the object represented by the icon (an old
    Japanese flag, or a German Black Cross, but you could as well put any symbol
    or flag of any country or army).

    I would much better like then to have cows or sheeps to count! Do we need to
    encode a cow or a sheep? And if so, why would we given it some numerical
    property? Every object has an implicit count of 1, and can be used for tally
    counts. What makes a character is some level of abstraction, such as the
    stroke for a rod, allowing representation on various supports and drawing
    with various tools, without having to represent the object exactly as it
    looks like.

    So for counting cows, we have the character "A" or any one of its ancestors.
    If we admitted flags as tally characters, then we would need to encode every
    object or concept which has a graphic symbol/icon associated to it. And then
    not only flags, but also the infinite heraldic symbols, and not only
    currency symbols but also the effective coins and bills (including color

    This archive was generated by hypermail 2.1.5 : Thu Mar 01 2007 - 06:39:20 CST