Re: Quick Question About Korean Input Methods

From: Charlie Ruland ☘ (ruland@luckymail.com)
Date: Tue Jan 05 2010 - 15:50:44 CST

  • Next message: Karl Pentzlin: "Re: Latin-script keyboard layout (was RE: Quick Question About Korean Input Methods)"

    Martin J. Dürst wrote:
    > Along the lines of Charlie's explanation below, I implemented a Korean
    > input method more than ten years ago in the application framework ET++.
    > (I gave a presentation about that at the 6th Unicode Implementers
    > Workshop in San Jose in 1994.)
    >
    > If you for example typed the letters "hana", then at the point you had
    > typed "han", it would all be one syllable, with the "n" being the final
    > consonant. Then if you typed another "a", the "n" you had typed before
    > would be borrowed back from the "han" syllable, that syllable therefore
    > changing into "ha", resulting in overall two syllables, "ha" and "na".
    >
    This is precisely the way Microsoft’s “Korean Input System (IME 2002)”
    works. Using Dubeolsik input mode on a QWERTY keyboard, ‘gks’ yields 한
    (han), ‘gksk’ 하나 (ha-na), ‘gksdk’ 한아 (han-a) etc.

    BTW, if you want to test this without installing this (or any other
    Korean) IME you may try the ‘Dubeolsik/QWERTY Converter’ at
    <http://lemon-factory.net/haconv>.

    Ganz herzliche Grüße ins Land der aufgehenden Sonne 日本,
    Charlie
    >
    > Overall, it looked somewhat funny, but I think it did the job. But I
    > never had it tested with native users.
    >
    > It worked only when typing text continuously (i.e. it wasn't possible to
    > 'steal' a final consonant from a syllable by placing the caret just
    > after that syllable). But it worked even if the application wasn't
    > prepared at all for such stuff. It did that by simply sending a
    > backspace character and then the two new syllables to the application.
    >
    > Regards, Martin.
    >
    > On 2010/01/03 2:21, Charlie Ruland ☘ wrote:
    >> Ed,
    >>
    >> please note that in modern Korean it is usually unnecessary to
    >> ‘finalize’ a syllable in order to distinguish between initials and
    >> finals. The reason for this is, of course, their distribution: any
    >> consonant immediately followed by a vowel/medial is an initial: there
    >> are no initial consonant clusters. Any consonant not immediately
    >> followed by a vowel/medial is a final.
    >> Please also note for the above that the ‘zero’ initial ㅇ is just like
    >> any other initial that has to be input, and that ‘strong’ obstruents
    >> that are written like double consonants (‘geminates’) have their own key
    >> combination (SHIFT+‘weak’ obstruent) and count as simple initials.
    >>
    >> Charlie
    >>
    >>
    >> Ed Trager wrote:
    >>> Hi, again everyone!
    >>>
    >>> Dreiheller, Albrecht wrote:
    >>>
    >>>> I would use<Space><Backspace>.
    >>>> But I'm not Korean, so there might be a shorter way.
    >>>
    >>> Yes, that's what I did too. But<Space><Backspace> is surely too
    >>> *slow* to be practical for really typing Korean, n-est-ce pas?
    >>>
    >>> On Fri, Jan 1, 2010 at 3:11 PM, Charlie Ruland ☘<ruland@luckymail.com>
    >>> wrote:
    >>>> For Hangeul I use Microsft’s “Korean Input System (IME 2002)”, and
    >>>> what I do to ‘finalize’ an open syllable followed by an initial
    >>>> consonant is press the ESC key. (I have no idea what the officially
    >>>> recommended key is.)
    >>>
    >>> I did not think of the ESCAPE key. But ESCAPE also is not
    >>> conveniently located on the keyboard. So using ESCAPE also will be
    >>> relatively *slow*.
    >>>
    >>> I asked the question because I am working on writing some input method
    >>> software ... Checking both Microsoft's and Apple's documentation on
    >>> Korean Input Methods, I don't find anything describing which
    >>> "official" key is to be used to "finalize" such syllables ...
    >>>
    >>> - Ed

    -- 
    Charlie • 查理 • चार्ली • Чарли • تشارلي
    チャーリー • 찰리 • Τσάρλι • צ׳ארלי
    oṃ āḥ hūṃ
    ॐ आः हूँ
    


    This archive was generated by hypermail 2.1.5 : Tue Jan 05 2010 - 15:56:11 CST