Re: kurdish sorani

From: Andries Brouwer (aebr@win.tue.nl)
Date: Tue Aug 29 2006 - 07:34:32 CDT

  • Next message: Andreas Prilop: "Re: kurdish sorani"

    On Mon, Aug 28, 2006 at 08:21:06PM -0700, John Hudson wrote:
    > Andries Brouwer wrote:
    >
    >> I think Uighur uses U+06D5 and U+06BE, but the shaping behaviour
    >> of Kurdish H seem to differ from that of Uighur H.
    >
    > There is currently uncertainty over the correct encoding of the Uighur h.
    > Although U+06BE is the obvious choice based on its shaping, at least some
    > Uighur text seems to use the Arabic heh U+0647, which requires dedicated
    > Uighur fonts or OT language system tag support.
    >
    > The Chinese draft standard for Uighur specifies U+06BE, which seems to me a
    > sensible choice.
    >
    > Note that the preferred Uighur glyph shapes may vary slightly from those
    > used for U+06BE in e.g. Urdu, but the differences are minimal.
    >
    > John Hudson

    Now that I look at Uyghur, it seems that U+06BE is no good for
    Uyghur either, so both Kurdish and Uyghur seem to need a new
    code point.

    As you say, the preferred glyph shapes are very different,
    but one might call that a matter of font. Worse is that
    the shaping behaviour is different.

    Arabic has 4 glyphs: isolated, initial, medial, final heh
    (in fact medial heh has two common shapes, but that may be
    a matter of font).

    Now both Kurdish E and Uyghur E need an isolated form and a
    final form, and they both use isolated heh and final heh.
    Let us call this behaviour U+06d5. No problem there.

    Both Kurdish H and Uyghur H use initial and medial heh,
    but the pattern is different:

    Position: Isol Init Med Final
    Kurdish H: init init med init
    Uyghur H: init init med med
    and
    Urdu: isol init init isol

    If the Urdu behaviour is called U+06be, then that is no good
    for Kurdish and Uyghur. Not only are the shapes rather different,
    but the distribution of the two shapes over the four positions
    differs.

    Andries



    This archive was generated by hypermail 2.1.5 : Tue Aug 29 2006 - 07:40:04 CDT