Re: Irish dotless I (was: Languages with letters that always take diacriticals

From: Michael Everson (everson@evertype.com)
Date: Thu Mar 18 2004 - 16:28:41 EST

  • Next message: Asmus Freytag: "Re: help needed with adding new character"

    At 14:32 -0600 2004-03-18, Brian wrote:

    > > Well, unless your spelling-checker author is bright enough (as is very
    > > likely) to handle both dot-convention and h-convention spellings.
    > > These are not intrinsically tied to Uncial vs. Antiqua font styles,
    > > though; one can write perfectly good Irish in Antiqua style and still
    > > use dotted consonants.
    >
    >That's my point. The differences between Roman and "Gaelic" orthographies as a
    >whole are not intrinsically tied to font styles. Although Michael insists that
    >it is,

    No, I don't. You can write "dóig*" in old
    orthography or "dóigh" 'likelihood' in modern
    orthography. The orthographies are different,
    just as "colour" and "color" are different. You
    can display "dóig*" and "dóigh" in a Roman or a
    Gaelic font style. The "d" is 0064, the "ó" is
    006F 00301 (or the equivalent 00F3), the "i" is
    0069, the "g*" is 0067 0307 (or the equivalent

    The "i" is not equivalent to 0131. It is a
    spelling error to use 0131 in Irish, and it will
    always be.

    >he also acknowledges that current spell checkers
    >only work with the modern (Roman) orthography
    >and that there are no spell checkers that work
    >with
    >the "older" orthography.

    Because no one needs one, and no one has made a
    corpus of texts in that orthography available. If
    someone makes one available, it *MUST* encode "i"
    as 0069. If it does not, then it will violate the
    standard set down in ISO/IEC 8859-14, which does
    not contain 0131.

    >Use of the dot-convention of marking lenition
    >already necessitates that different characters
    >must be used to represent the same phonological
    >segments that would otherwise be represented by
    ><cons>+"h", which would itself confound existing
    >spell checkers.

    That is one of the things which makes them different orthographies.

    >In this context, and if it's true that a spell checker could, in theory, be
    >programmed to handle parallel encoding conventions, then why shouldn't Irish
    >language "traditionalists" encode the i with a LATIN SMALL LETTER DOTLESS I
    >such as <0131>?

    Because that would be a spelling error. The
    letter "í" is the long form of "i". It is encoded
    0069 0301 (or its equivalent 00E9). It would also
    be a spelling error to encode "í" with 0131.

    Those are the facts. It is not a matter for dispute.

    -- 
    Michael Everson * * Everson Typography *  * http://www.evertype.com
    


    This archive was generated by hypermail 2.1.5 : Thu Mar 18 2004 - 17:12:41 EST