> It is true that most Japanese people have never read old Japanese
> books (such as Story of Genji, Hyakunin-issyu, Houjouki, and so on)
> in old Japanese handwriting, nor Kanbun (writting of Confusius,
> Lao-tse, and so on) in original typeface.
I was not referring to those only, but to all books printed before the
1960s (and some even later, like South-American Japanese publications).
By Kanbun, I mean the school curriculum that every Japanese schoolchild
goes through, which comprises not only Confucius and Lao-Tse but also
Chinese poetry written in both China and Japan. Now I know that most
people may be sleeping during many of these lessons or have forgotten
them, but one thing they will have learnt is that there is more to
Han characters (kanji) than just the standard decided by the Monbusho
in the 60s, and that you have to reckon with a certain variability
from time to time.
> Then what conclusion
> comes from this fact?
The simple conclusion that perfect monbusho-standard typesetting for
chinese-japanese textbooks at the xterm level is not a top priority,
given the poor condition that i18n as a whole is still in.
If the Japanese had wanted it to be easy to process CJK languages at the
same time smoothely by computer, they wouldn't have introduced those
monbusho reforms. Now I think it is fair that they must be patient with
the limitations that result from this. Not a high price anyway. Maybe I
shouldn't bring this political aspect into the discussion, but I think it
is needed to explain why this subject is dealt in such a heated manner
all the time.
> Do you insist Japanese people must learn the Chinese version of
> U+76F4?
No, I just say that this difficulty of displaying japanese+chinese
schoolbooks in XTerm is not a major concern for linux i18n. Moreover,
I would suspect that most Japanese people will easily recognize it in
a context where a certain tolerance to glyph variability is expected,
such as a C+J terminal text.
> Again, who pay the cost of education?
As B. Shaw said: "The reasonable man adapts to the circumstances. The
unreasonable man tries to adapt the circumstances to himself. All
progress comes from the unreasonable man."
If Japanese culture politicians wanted to revise a part of the reform of
the 50s and reunify at least some of the Han characters (as has been
suggested by many people in CJK countries, including Korean president Kim
Yongsam), the education costs would be even lower than those of the reform
of the 50s. I wonder why noone in the Japanese information processing
field was unreasonable enough to propose that.
> 1st level: we need different fonts with different glyphs for CJK,
> like -misc-fixed-medium-r-normal-{ja,ko,zh}-*. In this level,
> XTerm can use one of them for its whole life of process. For
> localization purpose (assume Japanese people read/write Japanese
> text only), this is enough.
I'd say that even these fonts should be complete unifonts that allow
also Chinese and Korean text to be displayed, even if not with the
perfect glyphs.
> 2nd level: we need some method to dynamically choose proper font
> according to the property of the text, not according to the
> users' locale environment. This is needed for further
> internationalization or multilingualization which allow foreign
> language to be displayed. (For example, a Chinese who want to
> learn Japanese or vise versa.)
I'd say "to be displayed perfectly according to all relevant national
standards".
> Ok, XFree86 will have different fonts for CJK (semi-1st-level-compliant)
> and XTerm will support language tag (2nd-level-compliant).
Yes. With this it seems that C+J textbooks can be perfectly displayed even
on xterm.
> > It is of course a bit unfortunate that we have to load different fonts for
> > all the tags.
> ...
>
> I agree. I imagine if Unicode were so designed that we don't have
> to discuss about such problem, i.e., Unicode were have a different
> policy of Han Unification. However, the fact is that at least
> Japanese people cannot read Chinese glyph and Unicode unified these
> characters. Fact is just a fact. We have to recognize the fact and
> manage this situation.
So are you proposing a way to avoid loading different fonts by
incorporating glyph variants in the fonts and selecting them on the
basis of language tags?
This sounds like a big subject for some standards group with possibilities
that go beyond what we are doing here.
-phm
-
Linux-UTF8: i18n of Linux on all levels
Archive: http://mail.nl.linux.org/lists/
