Re: Need for Localized Look of Shared Japanese/Chinese Gliphs??

From: John H. Jenkins (
Date: Tue Jul 14 1998 - 13:49:54 EDT

You're going to get a gazillion answers to this question, but--

On 7/14/98 9:03 AM, Marco Mussini (

>in UNICODE many glyphs (for example for kanjis) which are present both
>in Japanese and in Chinese, are "shared", i.e., they are assigned the
>same codepoint in the Unicode allocation, in order not to have redundant

The usual term is "unified."

>THe question is, those glyphs that were defined as "shared" for that
>reason, are similar enough, in their actual expected onscreen
>representation in the two languages, so that a single universal Unicode
>font may display them identical, or do you have any special example of a
>kanji that has the same fundamental appearance in Chinese and Japanese,
>but to be really well-localized it needs to be drawn in a slightly
>different way in Japanese than in Chinese, so a single font will never
>be enough, and we'll need a "universal Unicode font with the shared
>kanjis specialized with Japanese calligraphy" as well as a "universal
>Unicode font with the shared kanjis specialized with Chinese

BTW, I tend to find it interesting to hear East Asian ideographs referred
to as "kanji," since that's a Japanese word. But I digress.

First of all, it is utterly FALSE that there is a language-based
difference in "Japanese" and "Chinese" shapes for ideographs. (Sorry,
but this is a pet peeve of mine.) The difference is between Japanese
typography, mainland Chinese typography, and traditional Chinese
typography (e.g., Taiwan). There are lots of characters which will be
drawn one way in the PRC and another way on Taiwan, even though both are

The protoypical example is the "bone" radical, U+9AA8, which is drawn one
way in Japan/Taiwan and another way in the PRC.

Other examples may be found by perusing the on-line Unihan databse,

>In this case, to correctly *display* a Unicode string, it
>is still necessary to have locale information.

This is true.

>This is bad, what Unicode
>seems good for is to at least display the mixed-character-set text with
>a single font with 65,536 glyphs, rather than having to install several
>fonts and switching them on the fly.

Unicode isn't intended to provide full, ideal, typographic support in and
of itself. It's intended for "plain text." For plain text purposes,
most Japanese readers will simply use a Japanese font even for Chinese
text and be perfectly happy.

Moreover, "minimal legibility" is still possible without changing fonts
-- a Japanese reader will recognize and understand the characters written
in the "wrong" non-Japanese way.

Similarly, a font designed for German will look somewhat different from a
font designed for English. Nobody's going to insist, however, on
separating u-umlaut into a German and an English clone to support that in
plain text.

>By the way, I don't think that the kanjis that are clearly "shareable"
>between Chinese and Japanese do really have a vastly different drawing.

Japanese and Chinese think so. (Britons and Americans tend to find one
another's spellings of their common language bizarre, too.)

>However if you have an example of a kanji "shared" and with different
>appearance in the two languages, please describe it. I am a student of
>Japanese language and I can understand at least the japanese half of the
>example. But please, write it down in romaji!
>Also, does this sharing scenario happen also with Korean and other pairs
>of languages or is it only for Chinese and Japanese?

Again, it isn't true for "pairs of languages." It's true for different
national typographical traditions, which are partly language-based but
partly not. Taiwan and the PRC speak a common set of languages, all
called "Chinese."

And yes, it's also true for Korean.

John H. Jenkins

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:40 EDT