RE: Variant positions of combining marks

From: Ernest Cline (
Date: Sat Apr 10 2004 - 11:33:35 EDT

  • Next message: John Hudson: "Re: Variant positions of combining marks"

    > [Original Message]
    > From: Peter Kirk <>
    > To: Unicode List <>
    > Date: 4/10/2004 12:47:45 PM
    > Subject: Variant positions of combining marks
    > Does anyone know of any cases where the same combining mark is combined
    > with the same base character in two or more different positions, and
    > where the positioning is semantically significant? How does Unicode deal
    > with this situation? Would variation selectors be appropriate in this
    > situation, or what other mechanism might be proposed?
    > I am looking for examples other than Hebrew vav and holam, because
    > I am looking for precedents for dealing with this situation elsewhere in
    > Unicode. I am also excluding distinctions like DOT ABOVE and DOT
    > BELOW, where the marks are treated as different characters although
    > the glyphs may actually be the same.

    Since the fixed position of the holam variant under discussion is not the
    same as that of holam, a variation selector would definitely not be
    appropriate. Holam has canonical combining class 19, and as such
    any variant identified by a variation selector would be of that class and
    should therefore place its mark in the same fixed position as holam.

    The shin dot and sin dot pair applied to the letter shin is probably the
    closest analogy. Same base character, with identical looking marks
    distinguished by position only and having different fixed position
    canonical combining classes. That suggests that if the holam variant
    that is under discussion is accepted for Unicode, then it should be
    encoded as a separate character with a fixed position combining
    class other than 19. If one of the other fixed position classes used by
    Hebrew (Hebrew uses the range 10-26) is not identical to the position
    of this variant, then a new class will need to be added, perhaps class
    37 as that is the closest unused fixed position canonical combining
    class to the existing Hebrew classes.

    This archive was generated by hypermail 2.1.5 : Sat Apr 10 2004 - 12:07:40 EDT