RE: Extending the semantics of ZWJ and ZWNJ for Indic scripts

From: Andy White (
Date: Mon Nov 18 2002 - 21:34:44 EST

  • Next message: Kenneth Whistler: "Re: The result of the Plane 14 tag characters review"

    It has occurred to me that if we allow the 'primary function of ZWJ' to
    apply here, it may be better to encode Ra_Jophola as Ra ZWJ Ya.
    And it would then follow on that Vowel_A_zophola_AA can be encoded as
    Vowel_A ZWJ Ya, as suggested by Apurva.

    In Indic scripts, a Virama sign (commonly known as a Halant) applies to
    the consonant it is attached and as a side effect, may alter a
    proceeding character.
    It is not a control character.
    A Virama cancels the inherent vowel of the consonant that it attaches
    The isolated sequence 'Virama Ya' is meaningless.
    Virama was not invented to control the formation of a proceeding
    character and never should be.

    The sequence Vowel_A Virama Ya VowelSign_A (as in the Indic FAQ) is
    meaningless - unless you consider Virama as a control character - but is
    It is a real character with the real semantics as stated above.
    The Unicode consortium had no right to give it new semantics without
    consulting with the Indic community at large.
    How would it be if the Unicode consortium were to give the underscore
    the additional semantic of letter joining e.g. ffi.ligature can now be
    encoded as f_f_i ?

    The point of the matter is that to encode Vowel_A_zophola_AA as Vowel_A
    Virama Ya VowelSign_A is illogical as the Virama is attached to the
    Vowel_A which has no logical semantic at all.


    This archive was generated by hypermail 2.1.5 : Mon Nov 18 2002 - 22:26:14 EST