Re: Arabic letters separated by markup

From: John Hudson (
Date: Tue Jun 14 2005 - 01:33:53 CDT

  • Next message: Benjamin Kite: "Unihan and U+939D"

    Erik van der Poel wrote:

    > The OpenType spec appears to include fairly advanced mechanisms (see
    > e.g. contextual substitution), but there may not be very many fonts that
    > include ligature component info suitable for coloring in the way we have
    > been discussing. I'm guessing that the obligatory ligatures normally
    > consist of single glyphs (not multiple glyphs with positioning info).

    In OpenType terminology, that is exactly what a ligature is: a single glyph representing
    multiple characters, accessed through a ligature lookup type in the GSUB table.

    > There probably aren't very many system libraries and applications that
    > support ligature component coloring either. Maybe someone can mention
    > the names of fonts, system libraries and/or applications that actually
    > support this. Adobe, Apple, Microsoft?

    None. This is what I have been saying. The understanding of ligature in current font
    technologies is that of a single glyph representing multiple characters, not simply of two
    text elements, however encoded or rendered, that happen to appear joined in some way
    visually. There is no software that I'm aware of that can support ligature component
    colouring if one is using the term ligature in this technical sense. As noted earlier in
    this discussion, it is possible to render something that has the same visual appearance as
    a ligature but that actually uses more than one glyph, in which case the different glyphs
    making up this typeform* can be independently coloured easily enough in any application,
    so long as each glyph corresponds to one underlying character. Applications apply
    selection for colouring purposes, as for most other purposes, to characters, not glyphs.

    John Hudson

    * Typeform is a handy term adopted in the Dept. of Typography and Visual Communication at
    the University of Reading, which I've found useful in many discussions. It refers to a
    purely visual unit of rendered type, without reference to how it is rendered; indeed, one
    most often will not know how a particular typeform is rendered simply by looking at it. A
    letter with diacritic, for example , may be rendered as a single glyph or as two glyphs
    (it may also, of course be encoded as one or two characters). Similarly, in the current
    discussion, a lam_alif 'ligature' typeform may be rendered as a single glyph, i.e. a true
    ligature in the technical sense, or as two glyphs.

    Tiro Typeworks
    Vancouver, BC
    Currently reading:
    Truth and tolerance, by Benedict XVI, Cardinal Ratzinger as was
    War (revised edition), by Gwynne Dyer

    This archive was generated by hypermail 2.1.5 : Tue Jun 14 2005 - 01:35:26 CDT