About language-env
==================

language-env is a package to construct native language environment 
easily.  This environment construction is done by making or modifying
dot-files.  So each user has to construct his/her environment
by invoking set-language-env.

Dot-files means files in users' home directories whose names begin 
with dot (or period).

Invoking set-language-env inserts various settings for native
language enabling surrounded by "---- language-env" comments into
your dot files.  These settings are added at the last of your
existing dot files at the first time you invoke set-language-env.
You can add settings before or after the added region and
you can modify the added settings.  When you again invoke 
set-language-env (for example when language-env is updated), the 
old settings surrounded by "---- language-env" are replaced by the 
new settings.  Thus if you have modified the added settings, this
modification is lost.  Settings outside the surrounded region 
are not affected.  Don't modify the "---- language-env" lines.


How You Can Read Japanese Text
==============================

Some of you may read this document though your native language
is Japanese, because you cannot display Japanese.  Here is a small
HOWTO to display Japanese.

(1) Install "kon2", "konfont", and "lv" packages.
(2) Type "kon" to enable Japanese character on console.
(3) Use "lv" to view Japanese text files.


About Codeset
=============

The following codesets can express Japanese:
 * ISO-2022-JP (aka JIS code; used for mail and so on)
 * EUC-JP (Japanese EUC; usually used for Unix system)
 * SHIFT-JIS (aka MS Kanji Code; default and only code for Windows 
   and Macintosh and therefore the most popular codeset in Japan)
 * Unicode (UTF-8 is the most popular; not widely used yet)
This package adopts EUC-JP, according to "Policy on Japanese
Locale on Linux" Version 1.03, 
http://www.linux.or.jp/JF/JFdocs/Japanese-Locale-Policy.txt.

Since version 0.30, ja_JP.UTF-8 is used within "uxterm" (experimental).

Packages Needed for Japanese Environment
========================================

Though language-env establishes Japanese (and other languages) environment 
with modifying personal dot-files, it does not install additional packages 
needed for Japanese environment.  Now these needed packages are explained.


Mandatory
---------

All of you will need "locales" package.  Please add a line of
"ja_JP.EUC-JP EUC-JP" and invoke /etc/locale.gen to generate
ja_JP.eucJP locale.  Or, invoke "dpkg-reconfigure locales" in
root user and choose "ja_JP.EUC-JP" line.  Note that "locale-ja"
package is not needed any more.


Display
-------

At first, an environment has to be prepared where Japanese character 
or Kanji can be displayed. Install the following package:
 * 'kon2' for ordinal console
 * 'jfbterm' for frame buffer console
 * 'kterm' or 'rxvt-ml' for X Window System
Then invoke
 * 'kon' in ordinal console
 * 'jfbterm' in frame buffer console
 * 'kterm' or 'krxvt' in X Window System
You also need 'xfonts-cjk' package for X.

Besides 'kterm' and 'krxvt', there are several Japanese-capable
X terminal emulators such as 'mlterm', 'gnome-terminal', 'eterm',
'kwterm' in 'wterm-ml' package, and 'katerm' in 'aterm-ml' package.
However, you cannot input Japanese in 'kwterm' and 'katerm'
(only displaying).

Next, install Japanese-capable pager package 'lv' or 'jless'.
They can automatically distinguish codesets and properly display
Japanese text file written in various codesets.

Some window manager can display Japanese at window title and
others cannot.  'asclassic', 'qvwm', 'icewm', 'blackbox', 'sawfish,
and so on can without any additional settings.  'wmaker', 'ctwm', and 
so on can with additional settings.  However, this version of language-env
does not set anything on window managers.

# Of course environmental variable 'LANG' has to be set properly
# for all window managers.  This is well managed by language-env and
# you don't need to think about it.

If you use wmaker, invoke 'wsetfont japanese'.  Then modify
from 'MultiByteText = NO' to 'MultiByteText = YES' in
~/GNUstep/Defaults/WMGLOBAL and ~/GNUstep/Defaults/WindowMaker.


Input
-----

You need a Kana-Kanji conversion software to input Japanes.
You can choose one of the following softwares:
 * Canna : install 'canna'
 * Wnn : install 'freewnn-jserver'
 * SKK : install 'ddskk' and ('dbskkd-cdb' or 'skkserv')
   ('skk' is obsolete.  Use 'ddskk' instead.)
You can buy prorietery softwares with better convert efficiency but 
language-env does not support them.

There are two ways of connection between these conversion software
and softwares in which you'd like to input Japanese.  In non-X
environment, direct connection between conversion softwares and 
clients is used because there are no standard protocol.

   Conversion Softwares   ---->   Clients for each conversion software
   (Canna, Wnn, SKK)              (emacs20, mule2, jvim-canna,...)

In this way, you can use softwares which has faculty to connect
one (or more) of conversion softwares.

On the other hand, XIM protocol is the standard for X11R6.  It is
usual to use intermediate softwares which convert protocol of 
Japanese conversion softwares and XIM protocol.

   Conversion softwares  -->  Intermediate  -->  XIM Clients
   (Canna, Wnn, SKK)          (kinput2 etc)

Note that emacs20/emacs21/xemacs21 and so on connect directly
with conversion softwares even in X Window System environment.


At first you have to prepare connection between these conversion softwares
and each software you use.

If you'd like to use Canna or Wnn in emacs20 or emacs21, you will
need 'egg' package.  If you'd like to use Canna with 'egg' package,
you will need to edit /etc/init.d/canna file.  Details are explained
later.

Many softwares on X Window System use Kinput2 or XIM protocol.
You need
 * 'kinput2-canna' for Canna
 * 'kinput2-wnn' for Wnn
 * 'skkinput' for SKK
which connect between these conversion software and XIM protocol.
These softwares have to be running.  Language-env sets in .xsession
to invoke one of them.

Language-env sets .Xresource to use shift-space for invoking 
Japanese-conversion mode.


Text Processing
---------------

You can use
 * grep-ja
 * sed-ja
 * gawk-ja
for text processing which are multibyte character-sensible.  Debian 
package for Japanese version of perl is not available.


Others
------

The following are mail readers which can handle Japanese
 * mew
 * wl
 * im
 * sylpheed

The following are web browsers which can display Japanese
 * netscape with netscape-ja-resource-*
 * mozilla
 * galeon
 * lynx
 * w3m
 * w3mmee
 * w3 on emacs

The following are editors which can handle Japanese
 * emacs20
 * emacs21
 * xemacs21-mule
 * xemacs21-mule-canna-wnn
 * jed-ja and jed-canna
 * jvim-canna
 * yudit
 * gedit
 * dedit
 * vim
 * gvim (vim-gtk)
Note: read the following for emacs.

Softwares of -canna version can directly connect with canna server.
If you don't use X Window System and you don't like emacsen, you
can input Japanese with -canna softwares.

You can install many Japanese-related package using japanese 'task'
by using 'tasksel' command in 'tasksel' package, though you will be
forced to install many softwares.


Note on .emacs
==============

The following table shows supported methods with emacsen.

----------------------------
emacsen    Canna   Wnn    SKK
----------------------------
emacs20    egg     egg    OK
emacs21    egg     egg    OK
xemacs21   OK      OK     OK
----------------------------

  Note: 'Emacsen' is a generic name for emacs.

Debian GNU/Linux 3.0 (Woody) doesn't have mule2, emacs20-dl, and
xemacs20 packages.  language-env doesn't support these packages.
Please use emacs20, emacs21, or xemacs21 packages instead.

If you want to use emacs20 or emacs21 with Canna or Wnn, you will
need 'egg' package.

If you want to use Canna with 'egg' package, you have to edit
/etc/init.d/canna file like following:

   CANNA_SERVER_OPT="-inet -u canna"

(pease add '-inet'.)  Then, please restart Canna server by
typing "/etc/init.d/canna restart".

There are two methods to use SKK.
 - direct use
 - via dictionary server
Language-env supports the latter method only.  


Note on .xsession
=================

Behavior is modified by shell variables $WINDOW_MANAGER, 
$TERMINAL_EMULATOR, $PASS_THROUGH.  Read .xsession for detail.

If gnome-session is installed, gnome-session will be invoked
regardless of these variables.


Contact
=======

Tomohiro KUBOTA <kubota@debian.org>
http://www.debian.or.jp/~kubota/linuxwork/language-env.html


Copyright
=========

See /usr/share/doc/language-env/copyright.

---
Tomohiro KUBOTA <kubota@debian.org>  Mon, 18 Oct 1999