Re: stacking diacritcs and fullwidth characters

Date: Wed Jan 30 2008 - 04:56:45 CST

  • Next message: "Re: stacking diacritcs and fullwidth characters"

    Quoting James Kass <>:

    > John Knightley wrote,
    >> thank-you, GPOS certainly manages to centre the dots correctly-at
    >> least in some applications.
    >> U+0332 combining low line however needs to be half or full width
    >> depending upon the width of the character it is combined with.
    >> It would appear from comparng that Code2000 and SIL Doulos use some
    >> other mechanism to produce stacking.
    > Code2000 uses OpenType GPOS (glyph positioning) tables to stack diacritics
    > and to reasonably place diacritics in relation to their base characters. The
    > SIL Doulos font most probably uses the same method.
    Dear James,

    My apologies - I shoould rather say that Code2000 uses advanced GPOS
    to produce stacking, and that I have not been able to do the same.

    >> For some reason - the stacking feature, by mark to ligature does
    >> not seem to work. This however maybe due to other factors.
    > Can you post a graphic example of the notation you desire? It
    > is said that a picture is worth a thousand words, and this might
    > help us understand the desired behaviour.

    The attached sample.png is simple example, and sample2.png slightly
    more complicated. (I have added a few dots and lines, so that in a
    few notes one can see main variations.)

    Notes use the numbers 1 to 7; dots above or belowso a higher or lower
    octave; a line below halves the duration of the note.

    With half width numbers this is no problem.


    > The fullwidth presentation forms of Latin characters (and western digits)
    > may present an interesting problem. I'd have to check, but I'm pretty
    > sure that I've made no allowances for any of them using GPOS and
    > combining marks. I'm also not sure if a rendering engine will attempt
    > to shape characters from that range, but it would make for an interesting
    > test. (Even the ASCII digits -- I had not considered that anyone would need
    > to put combining marks with them and so did not make any effort to
    > assign what are called "anchor points" to them.)
    > Best regards,
    > James Kass

    This message sent through Virus Free Email


    This archive was generated by hypermail 2.1.5 : Wed Jan 30 2008 - 07:47:52 CST