Re: The Bidi Class of the Khmer Symbols U+17F0 to U+17F9

From: Asmus Freytag (
Date: Wed Apr 28 2004 - 21:36:29 EDT

  • Next message: Mike Ayers: "RE: PUA as the Wild West [was: SSP default ignorable characters]"

    At 05:25 PM 4/28/2004, Kenneth Whistler wrote:

    > > So my question is, is this Bidi Class of these numbers correct?
    >As best we know.

    Actually the analysis we did is probably faulty.

    We use ON for symbols, because we want to allow generic symbols to appear
    as part of both RTL and LTR text. In the case of these characters, even if
    they are symbols, they are by definition part of Khmer text.

    Therefore, ideally, these should have been given a property of 'L'.

    On the other hand, these characters are not high frequency even in their
    intended domain. Therefore, it's actually not critical that they be given a
    particular bidi class, as long as the class they have works correctly in a
    Khmer context. ON works for that.

    Changing their assignment is disruptive and would make some implementations
    do things differently than others. It's better that all implementations
    have the same behvior. even if it requires the use of an occasional extra
    LRM for some obscure Khmer/Arabic document.


    > > Since their behavior when used with Khmer would be the same
    > > as if they had been given Bidi Class L, if it is an error, it is
    > > understandable how it could have escaped notice by the
    > > users of this script. If it isn't an error, could someone please
    > > explain why they are have class ON instead of class L?
    >Because they were represented to the UTC as being a bunch of
    >symbols, essentially.
    >Keep in mind that these are not at all normal digits. They are
    >exceedingly esoteric, and from the examples in the originating
    >documents it isn't clear how they lay out with other *Khmer*
    >characters, let alone any other script.
    >It is quite likely that there exists no document on the planet
    >which uses them in a bidirectional context, so your guess
    >is as good as mine what the original users of these divination
    >characters would have decided to do if faced with the task
    >of having to write them in an Arabic document.

    This archive was generated by hypermail 2.1.5 : Wed Apr 28 2004 - 22:15:55 EDT