Re: Arabic letters separated by markup

From: John Hudson (
Date: Tue Jun 14 2005 - 13:37:47 CDT

  • Next message: Tom Emerson: "Re: It is easy to predict the past."

    Erik van der Poel wrote:

    > Any application? I was replying to Mete Kural, who was probably
    > referring to Arabic. Are you aware of any application that can render
    > all the normal Arabic glyphs (including joining and ligatures) and also
    > color the components of a ligature-like typeform, e.g. lam-alef? If so,
    > which app(s), and which OpenType mechanism(s) would be used for that?

    Take MS Word for example. You can make selections at the character level and apply colour
    to the glyphs that represent the selected characters. So long as there is a one-to-one
    relationship between characters and glyphs, applying colour is easy (it is also presumably
    easy if there is a one-to-many relationship, i.e. if a single character is represented by
    two glyphs, both should be able to be coloured together when the underlying character is

    So a 'ligature-like' typeform that is actually composed of two glyphs, each representing
    one underlying character, should not pose a problem for colouring of the individual
    component characters, because although it *looks* like a ligature it is not different from
    any other sequence of two glyphs representing two characters. [In order to form the
    'ligature-like' typeform from the two glyphs, one would presumably use a two-step
    contextual alternates <calt> lookup, e.g. using VOLT syntax:

            lam.init -> lam.initlig
            lam.medi -> lam.medilig
            | alif.fina

            alif.fina -> alif.ligfina
            lam.initlig |
            lam.medilig |

    So the input glyph sequence /lam.init/alif.fina/ is rendered using
    /lam.initlig/alif.ligfina/, these two glyphs forming the 'ligature-like' typeform.

    Again, the fact that the typeform looks like a ligature is irrelevant: all you have is two
    glyphs in sequence corresponding to two characters. How you got to those particular two
    glyphs, the contextual substitutions, is irrelevant for colouring purposes. All that
    matters is that each selected character is represented by an independent glyph.

    I don't have a test font that handles the lam_alif typeform in this way, but I have plenty
    of fonts that use contextual substitutions in other ways, and am able to colour resulting
    glyphs so long as they correspond to individual characters.

    John Hudson

    * If two glyphs represent a single character, i.e. a character that has been decomposed at
    the glyph level, it should be trivial to colour both the resulting glyphs with the same
    colour. Much more problematic, of course, would be to colour one of the glyphs in a
    different colour from the other. This cannot be handled through selection colouring,
    because selecting the underlying character will always imply selecting both glyphs. This
    kind of colouring can only be done at a higher level based on GDEF property distinction
    between the two glyphs. This is what MS Word does when e.g. one decomposes the alif with
    hamza character at the glyph level to display as separate /alif/combining hamza/ glyphs.
    It is possible (globally in Word, using complex script mark display options) to
    independently colour the combining hamza because it is identified as a 'mark' in the GDEF

    Tiro Typeworks
    Vancouver, BC
    Currently reading:
    Truth and tolerance, by Benedict XVI, Cardinal Ratzinger as was
    An autobiography from the Jesuit underground, by William Weston SJ
    War (revised edition), by Gwynne Dyer

    This archive was generated by hypermail 2.1.5 : Tue Jun 14 2005 - 13:39:29 CDT