RE: Greek curled beta in Unicode code chart

From: Asmus Freytag (
Date: Tue Jul 05 2005 - 12:58:45 CDT

  • Next message: Kenneth Whistler: "Re: sub and superscripts"

    At 05:06 AM 7/3/2005, David Perry wrote:
    >Scripsit Alexej:
    > > I think, this the only reason, why the curled beta, U+03D0,
    > > was introduced into the Unicode in addition to standard beta with
    > > descender (AFAIK, this glyph form is completely unknown to mathematicians).

    Alexej is correct in that this form, unlike the forms for theta, etc, is
    not present in MathClass-6.txt or the ISO9573 mappings from scientific
    entity sets. Consequently it is probably also not be part of MathML 2.0, but
    I've not verified that separately.

    However, documenting mathematical use is challenging, and the proof that
    something is 'never' used in mathematics is generally impossible. Especially
    for forms that would have been available to hot-lead typesetters.

    >The various alternate Greek letter shapes (curly beta, curly rho, "script"
    >kappa, etc.) were put in Unicode because they were present in pre-Unicode
    >character sets. They really don't belong in Unicode because they do not
    >follow the encoding model that Unicode is based on (encoding characters not

    David, I think, overstates the case here. The kappa and rho are definitely in
    use for mathematics; while their original inclusion may have been triggered
    by their presence in one or the other legacy set, they would have been added
    when we did the extensions to make sure that mathematical usage was covered.

    >More important, these characters should be used with great caution, if at

    Their use in mathematical notation is fine and does not require any
    particular caution. Their use in *ordinary Greek text*, however, is
    clearly inappropriate. However, claiming that "they don't belong"
    in Unicode is like claiming Unicode should be limited to ordinary

    >The best way to get alternate letter shapes is to use advanced font
    >technologies such as AAT or OpenType that allow the display of alternate
    >glyphs without modifying the underlying Unicode values.

    This is the correct solution for getting the correct display of variant
    forms in ordinary text - especially where the need for the variant form
    can be predicted with accuracy from context.


    This archive was generated by hypermail 2.1.5 : Tue Jul 05 2005 - 12:59:43 CDT