Re: old Latin chars

From: James Kass (
Date: Tue Jan 08 2008 - 04:00:49 CST

  • Next message: Khaled Hosny: "Re: chairless hamza"

    Using GSUB (glyph substitution), the font's tables can provide
    the outline data from a pre-composed glyph for a Unicode string
    consisting of a base letter plus a mark.

    Using GPOS (glyph positioning), the font's tables can position that
    mark relative to the base letter and provide the system with the
    exact same outline data as it would have done using the pre-composed
    glyph mentioned above.

    GSUB vs GPOS is a matter of individual preference. There should
    be no rendering difference when the outline data the font provides
    is the same.

    Anyone using certain modern systems should be able to display base
    letters with marks properly. In a perfect world, it would not be
    necessary to put such anticipated combinations in the P.U.A. as
    pre-composed glyphs. In the real world, there is no need to use
    the P.U.A. to *encode* such sequences as unique "characters". There
    is, however, still a desire to be able to *display* such sequences
    properly. MUFI provides a method by which this can be accompished,
    even on some older systems.

    Reading through some of the MUFI documentation, it is clear that
    proper caveats about P.U.A. use are in place.

    Best regards,

    James Kass

    P.S. - Related to stacking marks on base letters, this page from
    Andrew West concerning tests made with Doulos SIL may be of

    This archive was generated by hypermail 2.1.5 : Tue Jan 08 2008 - 04:06:39 CST