Re: kurdish sorani

From: Andries Brouwer (
Date: Wed Aug 30 2006 - 05:19:15 CDT

  • Next message: Philippe Verdy: "Re: Unicode characters List instead of hexadecimal equivalent"

    On Tue, Aug 29, 2006 at 07:46:24PM -0700, John Hudson wrote:
    > Andries Brouwer wrote:
    >> Your image agrees with my description of Uyghur shaping. Good.
    >> Concerning Urdu shaping, ...
    >> I am a bit unhappy with the Urdu line on your picture.
    > This is basically a matter of type style. As you note, Urdu is normally
    > written and typeset in the nasta'liq style, ...
    > In the nasta'liq style, the do chashmi he is essentially the same letter
    > shape regardless of its context, and the connecting line is through the
    > centre, on the diagonal. Indeed, in Zaidi & Hassan's _An Introduction to
    > Urdu Script_ (Central Institute for Indian Languages, Mysore, 1997), the do
    > chashmi he is uniquely presented as a letter without allographs.
    > The question of how best to represent this letter in horizontal, neo-naskh
    > type styles in an interesting one, but the dominant convention that has
    > emerged is one which matches what I have shown in my illustration, which in
    > turn corresponds to the desired Uighur h shaping. This has led several
    > different parties, including the relevant Chinese standards body, to
    > conclude that U+06BE should be used for the Uighur h.
    > I believe this is correct not only because in the style of type preferred
    > for Uighur text this shaping of U+06BE is the norm, but also because Uighur
    > manuscripts exist in the nasta'liq style, and I believe these will be found
    > to follow the conventions of Urdu nata'liq in the shaping of the do chashmi
    > he therein used as the Uighur h. Unfortunately, the one manuscript text I
    > have images of does not seem to seem to feature the letter h.

    So, it seems in the end we have come to agree on all the facts,
    and the only disagreement is as to what conclusion follows from
    these facts.

    Arabic/Persian Heh has four shapes, and Kurdish H has two shapes
    (namely two from the four used by Heh), and Uyghur H has two shapes
    (namely two from the four used by Heh), and Urdu has one or two shapes
    (different from the four used by its `small' He).

    If one is forced to use for Uyghur H a code point that already exists,
    I can see that U+06BE, introduced for Urdu, is the best choice.
    But that is a hack. It is wrong, but the best one can do
    with today's standard.

    The reason that I call it wrong is that Uyghur shaping is well-defined,
    while Urdu in the most common font shape uses only a single form,
    and manuscripts and fonts that use two shapes vary in usage.

    The reasoning used to establish that Uyghur H can use the Urdu
    U+06BE since the latter is ill-defined anyway and can be used
    in the Uyghur way, can also be used to establish that Kurdish H
    can use the Urdu U+06BE. But Uyghur and Kurdish treatments seem
    to be well-defined and distinct.


    This archive was generated by hypermail 2.1.5 : Wed Aug 30 2006 - 05:23:34 CDT