Re: Decomposition vs Full decomposition?

From: Mark E. Shoulson (mark@kli.org)
Date: Wed Mar 16 2005 - 09:37:17 CST

  • Next message: Mark E. Shoulson: "Re: French accented characters - observations of problems"

    Peter Kirk wrote:

    > On 15/03/2005 21:28, Michael Everson wrote:
    >
    >> At 13:16 -0800 2005-03-15, Peter Constable wrote:
    >>
    >>> U+048A CYRILLIC CAPITAL LETTER SHORT I WITH TAIL
    >>
    >>
    >>
    >> The Cyrillic short thingy isn't a combining character. And it's not a
    >> breve. (You knew this, but others mightn't.)
    >
    >
    > The only problem with that is that there is no CYRILLIC CAPITAL LETTER
    > I WITH TAIL, because CYRILLIC CAPITAL LETTER I is one of the few
    > Cyrillic letters which is not modified with a tail - except when
    > combined with a breve. So, what we have is a precomposed character
    > which consists of an existing combining mark combined with a base
    > character which is used only with the combining mark. I wonder if this
    > is a unique situation? I think not, because there is an Arabic chair
    > character which is only used with a hamza. And that situation is also
    > problematic. Also I think U+0640 ARABIC TATWEEL is supposed to be used
    > only with combining marks. Anything else? In pointed Hebrew this is
    > true of the letter shin, but this is used without combining marks in
    > unpointed Hebrew.
    >
    Or even in fully-pointed texts in the specific case of the name Issachar
    יִשָּׂשכָר, according to the (pretty much universally-accepted) Ben-Asher
    pointing (Ben-Naftali has יִשְׁשָׂכָר). I've been reading up lately on other
    traditions, etc.

    The situation you point out reminds me of Ladusaw & Pullam's "heng"
    character, which has no serious meaning or use, but is listed mainly to
    give the "hooktop heng" letter, ɧ U+0267 LATIN SMALL LETTER HENG WITH
    HOOK, a name. Apparently the same logic is being applied by Unicode,
    since U+0267, by its name, would appear to be a letter Heng with a hook,
    but nowhere do we find a letter Heng.

    ~mark



    This archive was generated by hypermail 2.1.5 : Wed Mar 16 2005 - 09:38:54 CST