[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Hangul Jamo and wcwidth()



 > For combining marks, it is to some degree feasible to let applications
 > such as xterm generate them on-the-fly by simply overstriking two
 > glyphs. However, I don't see, how this would work with the the
 > conjoining Jamo. It was my understanding that you actually have to
 > hand-generate the Hangul glyphs, as the Jamo components are not located
 > at fixed geometrix positions within the glyph. If I set wcwidth() to 0
 > as you suggest, xterm would simply overstrike the glyphs, which is also
 > not too useful. If we fix something at wcwidth, I'd rather prefer to
 > set all Jamo to either 1 or 2.

'Hand-generating' is very exaggerated.  A common technique uses 10
variants per leading consonant, two to four per vowel, and two to five
per final (so in total about 300 glyphs) to generate (by simple
overstriking) quite pleasant glyphs for the Hangul block (11000 odd
glyphs).

It would be quite feasible to draw a set of conjoining Jamo glyphs
that will look readable when simple overstriking is used (assuming
that sequences are well-formed enough to always start with a
consonant).  I would even do the drawing if you don't mind having the
glyhs in the font :-)

 > You probably misunderstood: The keysyms are all in English of course.

Well, I meant: hanIM seems to react to the keysym 0x0061, not 0x0eb1
(Hangul_mieum), and I have in fact never seen a header file defining
these Hangul_* symbols.

 > Korean Unicode users will have to wait until we have UTF-8 locales
 > running and then their IM has to be made fit to produce UTF-8. Or
 > use cut&paste from the font repertoire files as the most universal
 > input method available in the mean time ...

Well, what do non-Korean Unicode users use to enter characters not on
the keyboard into their X-applications?  What I would like to see is
an open IM implementation for X that can be configured through mapping
tables and dictionaries like the Emacs quail system, and that could be
used for practically any language.  Is there such a beast?  Perhaps
freeWnn would fit the bill.  

Otfried


-
Linux-UTF8:   i18n of Linux on all levels
Archive:      http://mail.nl.linux.org/lists/