Re: French group separators, was Re: The character for 10**24 inJapanesenumbers (jo)

From: Jim Allan (
Date: Tue Jul 08 2003 - 09:55:16 EDT

  • Next message: John Cowan: "Re: UTF-8 to UTF-16LE"

    Tex Texin posted on use of U+2007 FIGURE SPACE for digit-grouping space:

    > Right. I was only thinking that if U+202F wasn't available it might be a
    > better choice than NBSP.

    I checked some common fonts which confirmed what I believed, that digits
    are normally equal in width to the lowercase letter _n_ or very close
    and that a normal space (and non-breaking space) is half that width.

    Accordingly U+00A0 (very, very rarely not available in a font) seems a
    better choice than U+202F.

    U+202F as a digit-grouping space would be double the width of a normal
    word-division space, separating the parts of a single number more than
    words are separated, equal to sentence division if one is double spacing
    following a sentence.

    Of course in justified text both normal space and non-breaking space
    usually expand in width to enable line-filling and might so the
    digit-grouping space might sometimes expand to a width greater than the
    digit-width in that font.

    But even in such a case the digit-grouping spaces would still not be
    greater in width than word breaking spaces in the same line.

    U+202F which is always a wide space would be generally less desireable
    than ordinary non-breaking U+00A0.

    Jim Allan

    This archive was generated by hypermail 2.1.5 : Tue Jul 08 2003 - 10:51:53 EDT