Maybe it makes a sence to make something like this in Plan9 (an analog
kbmap) for typing complex symbols like an hieroglyph ?
Your method is in essence what Microsoft's IME on Windows and various IMEs
on UNIX-likes (such as SCUM) use. However, an IME for inputting from a list
of over twenty thousand characters takes quite an effort to devise before
it can be practical and useful. Right now even display of CJK is not quite
fully supported on any existing FOSS platform (Ruby character display was
added to Firefox only somewhere after version 3). Non-integrated pieces of
FOSS with great capabilities do exist.
In case of (Simplified and Traditional) Chinese there apparently exist only
two successful IMEs out there: one is Microsoft's, the other belongs to a
Chinese company that has put lots of money and effort into developing the
software. I believe both support input by Pinyin romanization, although I
may be wrong. There's also Google's Pinyin IME which was involved in a
lawsuit with said Chinese company.
In case of Japanese an IME needs to support three writing systems at once,
firstly the two kana, and then transforming from kana to kanji. Abundance
of homonyms in Japanese as well as a certain writing strategy called ateji
(using kanji for phonetic value rather than semantic value) makes embedding
of a dictionary into the IME unavoidable. Good dictionaries for this
purpose don't come free--they must either be bought from professional
companies or compiled by people who intimately know the language,
preferably native speakers. This latter, I believe, is how IMEs on
UNIX-likes came to be. Anyhow, Japanese IMEs, too, rely on input based on a
romanization of the language. The actual number of distinct kanji required
for input of text at a high school literate level is around two
thousand--JLPT Level One roughly corresponds to that--but people, of
course, expect a much larger dictionary. Microsoft IME also provides
semantic aid by offering short descriptions of kanji so that people can
decide which corresponds to the meaning they want to convey. Although
unnecessary, it is a most welcome addition.
I don't know anything about Korean writing system or IMEs but since CJK
ideographs (most importantly Han characters) are involved similar
statements may apply.
Overall, there's no easy way that is light on financial and/or human
resources--the two types of resources are interchangeable, i.e. if you have
an active user base you may be able to avoid expenditure--to put CJK input
support into a UI, which is probably why Plan 9 doesn't have that at the
moment. It isn't a computer thing--it's a human thing. I might add porting
IMEs from some UNIX-like system is probably the best option (for those with
the technical prowess).
**********
DISTRACTION
While googling around for the existence of IMEs on Plan 9 I came across
this document from 1996 titled "Unicode: Writing in the Global Village:"
Despite these hurdles, Unicode may soon become the most common
multilingual character-coding system. Support for multiple-language use
is quickly growing. New operating systems—AT&T's Plan 9, Windows NT,
Novell's Netware 4.01 Directory Services, Sybase's Gain Momentum, and
Apple's Newton already support Unicode.
--
<http://www.nyu.edu/its/pubs/connect/archives/96fall/hargitaivillage.html>
It's funny how the author assumes display and input are the same thing
while they so greatly differ, input being times harder to implement.
--On Friday, September 11, 2009 15:29 +0400 Alexander Sychev
<[email protected]> wrote:
Hello!
Some time ago I wrote for inferno an analog of kbmap with an extention -
a possibility to print complex symbols via sequences of more basic
symbols.
I use it for typing by the russian translit.
Here is a piece of file for my kbmap:
<------------cut --------------->
1 45 0
1 46 'Ц
1 47 'В
1 48 'Б
1 49 'Н
1 50 'М
C цх 'ч
C Цх 'Ч
C сх 'ш
C Сх 'Ш
C сцх 'щ
C Сцх 'Щ
<------------cut--------------->
The latin symbols are mapped to russian when it is possible. Other
russian symbols are presented via sequences of mapped symbols, e.g.
russian symbol 'Ч' [ch] is presented like an sequence of 'ц' [c] и
'х' [h].
A sequence can be broken by pressing any non-symbol key.
There is at least one big disadvantage of this method - the input focus
can be changed, e.g. by mouse. In inferno I didn't resolve this problem,
because /dev/pointer can be opened only once.
Maybe it makes a sence to make something like this in Plan9 (an analog
kbmap) for typing complex symbols like an hieroglyph ?
On Fri, 11 Sep 2009 14:23:02 +0400, erik quanstrom
<[email protected]> wrote:
HI..everyone:
Is there some ways to input Simplified Chinese in plan 9 ? I
know plan 9 supports Unicode, so it is no questions for plan 9 to
display Simplified Chinese....... and i have seen some pictures on
Internet to prove it...so i have a question like that above...
I'm looking forward for the answer........... Thanks first......!!!!
the only way to input simplified chinese currently
is to use the general codepoint input method.
<compose> 'x' + four hex digits. on a pc compose =
<alt>. that's probably not what you're looking for.
i am not aware that anyone has written an input
method specificly for simplified chinese.
- erik
--
Best regards,
santucco
