> > This, however, is absurd - one of those 1,000,000 words is
> > "antidisestablishmentarianism", and there's a whole bunch half that long
> > longer. Show me the glyphs for them! This NEC thingy may make cute
> > stuff, but it would be useless for communication. Besides, does anyone
> > really believe that alphabetic writers would decide that they'd rather
> > thousands of glyphs? We're getting deeply fictional here...
> Not to mention the controversy there'd be about unifying the glyphs for
> "colour/color" etc.
From my point of view, encoding "hataraku" and "ugoku" each with a code
point of its own has a certain striking similarity to providing code points
for the words like color and colour. I could have brought up a more
completely parallel example, but do you have the fonts on your machine to
show the JIS appendix characters introduced since level 2? The ones I am
thinking about are probably in the 3.0 standard, and, if not, most likely in
extension B. But do you have the fonts?
Many Japanese people deny this, but the closest parallel Japanese has to
what English speakers call letters are the sub-characters English students
of Japanese call radicals. (Kana are more comparable to a phonetic alphabet,
and I avoid refering too casually to Chinese because I have no experience
there.) Encoding the character ugoku is very much like encoding a
spelled-out word in English. Hataraku was originally a variant writing of
ugoku. Now, hataraku means to work and ugoku means to move. The components
of ugoku are "omotai" (heavy, by loose translation) and "chikara". Hataraku
adds a component "hito" (person) on the left. (This is a loose description,
but I don't have time for a more formal effort.)
They have encoded the pre-composited characters separately here primarily
because trying to render on the fly was impossible until fairly recently.
Well, actually, there are other technical issues -- compositing on the fly
still will not produce very pleasing results. (But it could produce legible
I'm sure a lot of the participants on this list are familiar with these
issues. But I think it is worth looking at them again. That's why I brought
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:19 EDT