Re: ct, fj and blackletter ligatures

Date: Thu Nov 07 2002 - 17:57:04 EST

  • Next message: John Hudson: "Re: A .notdef glyph"

    On 11/07/2002 12:43:22 PM Thomas Lotze wrote:

    >> As for providing a
    >> notification dialog to say that the text contains < c, ZWJ, t > but
    >> that the font doesn't support it, there are no existing mechanisms to
    >> support that at present,
    >I don't understand this. Since a font doesn't "do" anything, but
    >software using the font does, one could write a rendering engine which
    >gives feedback

    Sure, if you're writing software that interprets OT lookups, you could do
    this. If you want to write your own code to process that state tables in
    AAT or Graphite fonts, you could do this. The software that currently do
    these thing and that most app developers are going to rely on do not.

    >That's no reason. Just because many people don't need a feature, or
    >maybe just don't care enough to complain, doesn't mean it shouldn't be
    >provided for those who do need it.

    If you want it, write to the vendor that creates whatever software you
    use, and ask them to support it. (And good luck!) But I really don't
    expect you're going to find the font / rendering industry as a whole to
    agree that this is something that needs to be supported across rendering
    systems; there are too many important issues waiting to be dealt with to
    worry about this.

    >I could well imagine that in a typesetting application, it would make
    >sense to be informed on whether a certain typographic feature can or
    >cannot be applied.

    In a typesetting application, you would be immediately informed: you see
    it in your proof, or you do not. If it's not there, you act accordingly.
    In typesetting, you probably don't want to send character data to your
    service bureau and assume that their rendering process will produce the
    same results. They may not have exactly the same typeface or match other
    aspects of rendering, let alone not get the ligature where you want.

    > Carefully checking a long document for whether, e.g.,
    >a certain pair of glyphs does form a ligature where it is supposed to is
    >tedious and error-prone if done by a human, but such routine tasks are
    >what computers are good at.

    But if you're an author wanting a ligature, you don't need to proof read
    the entire long document; you just enter one instance and check it, and
    you'll known right away what the results for the rest of the document will

    - Peter

    Peter Constable

    Non-Roman Script Initiative, SIL International
    7500 W. Camp Wisdom Rd., Dallas, TX 75236, USA
    Tel: +1 972 708 7485
    E-mail: <>

    This archive was generated by hypermail 2.1.5 : Thu Nov 07 2002 - 18:39:42 EST