Arabic Presentation Forms A vs. B

From: ryan77@canada.com
Date: Fri Feb 20 2004 - 00:19:24 EST

  • Next message: Ernst Tremel: "sorting feature for Arabic Kurdish"

    Greetings,

    I have developed an application which attempts to store
    Arabic in Presentation Form in a data file. I realize
    this is not the proper way to store Arabic, however,
    the application which accesses the data operates on
    hardware with very tight memory/speed constraints and
    it was decided that turning standard form into
    presentation form at run-time would be too costly.

    My conversion routine uses joining rules from FreeType
    1 and uses the substitution table resulting from the
    following perl script:

    http://mail.nl.linux.org/linux-utf8/2000-10/msg00057.html

    Recently I've discovered that I've been missing the two
    element Ligatures in Arabic Presentation Forms B. Upon
    further research at unicode.org I also realized that I
    missed the three element ligatures and word ligatures
    in Arabic Presentation Forms A.

    This leads me to a few questions:

    The Arabic data in question is for place names in a
    mapping product. So far, we have only received one
    complaint and it was a missed two element ligature from
    Arabic Presentation Forms B. Does this mean that
    the ligatures in Arabic Presentation Forms A do not
    apply to place names? Or did I just luck out?

    If it is necessary to convert to elements in Arabic
    Presentation Forms A, how would I determine the
    substitution rules for multi element ligatures?

    Thanks for your time.

    Cheers,
    Ryan

    This archive was generated by hypermail 2.1.5 : Fri Feb 20 2004 - 01:07:08 EST