I'm sorry that I forgot to mention World21 before.

> I'm sorry if I hadn't make it clear at my home page where you would
> be able to find a sentence like "Runtime font loading" which is
> already an _implemented_ feature. Why I released the font in
> readable format as wsfont is because I needed to make sure every
> glyph is correct.

I think you don't have to make sure that, because there is no need
to include Japanese character in kernel source. If it is loadable,
it is already good enough.

> >         http://www.asahi-net.or.jp/~pf5y-inue/world21-beta7.tgz
> >                 (unfortunatelly, currently for NetBSD-1.4.*)
> 
> Unfortunately I can't read Japanese text at all. :(

But I think you can read the source mentioned above. :)

> I wonder: is it a patch for wscons or an userland program? 

This is set of sources in both kernel and userland.

> Why hasn't it been included as part of the standard distribution? 

There are several reasons, I think.

One of the reasons is perhaps Inoue-san don't have spare time to
update it. (It is currently 1.4.* based).

Also, I think there are at least two things to be corrected.

- Currently, codeset handling is hardcoded in wdpyemul_world21.c,
  thus, newer codeset like UTF-8 cannot be easily supported.
  This source file should be separated by codeset, thus, newer 
  codeset can be added easily and (more importantly) dynamically by LKM.
- Current intefarce of putchar/putchar2 in wdpyvar.h takes font
  argument. But as Minoura-san already said in this mailing list,
  there are machines like NEC PC98 which can display larger character
  set like JIS X0208 by character generator (not by frame buffer).
  Thus, this interface should take character-set-ID rather than
  font as argument. 

But I haven't told these problems to Inoue-san, so these are not
Inoue-san's fault.

> Can such codesets and fonts be displayed _together_ on a screen?

Theoretically it can be. Although I'm not sure it currently can.
(The days I usually used World21 are very old.)
--
soda