gjiten Manual V2.0

Botond Botyanszki

    <boti@rocketmail.com>

   Copyright  1999 - 2001 Botond Botyanszki
     ___________________________________________Done.
________

   Table of Contents
   1. Introduction
   2. History
   3. Installation
   4. Dictionary files
   5. Usage

        5.1. Word dictionary usage
        5.2. Kanji dictionary usage
        5.3. Command Line
        5.4. KanjiPad

   6. Troubleshooting

        6.1. File errors
        6.2. Japanese input (mostly Kinput2)
        6.3. Window manager and Kinput2 problems
        6.4. Fonts and mojibake
        6.5. Pasting from Netscape
        6.6. Settings
        6.7. Bugs

   7. Future Improvements
   8. Feedback
   9. Credits
     _________________________________________________________

1. Introduction

   Gjiten is a Japanese dictionary program. It also has a kanji
   dictionary. Any combination of stroke number, radicals and
   search key can be used for kanji lookups. Gjiten requires a
   working X Input Method [e.g. kinput2] for Japanese input.

   This program is licensed under the terms of the GNU GPL, a
   copy of which you should have received with this package. See
   the file COPYING for more details.
     _________________________________________________________

2. History

   I wrote gjiten because I needed a free software dictionary for
   my Japanese studies and wasn't satisfied with xjdic. Learning
   all xjdic's keys and running it in the ugly kterm isn't my
   taste. The other reason is that I've been using so much
   software written by the Open Source Software community without
   any contributions, so I thought it was time to write something
   ;) Hopefully someone finds this little program useful. Older
   relesases (pre 2.0) used some code from xjdic written by Jim
   Breen, but this has been removed.
     _________________________________________________________

3. Installation

   Rolling your own from a tarball:
            % tar zxpvf gjiten-x.x.tar.gz
            % cd gjiten-x.x
            % ./configure
            % make
            % make install

   Or you can make a binary rpm the following way. You need to be
   root for this.
            % rpm -tb gjiten-x.x.tar.gz

   The rpm package should be created under
   /usr/src/RPM/RPMS/`uname -m`/ See rpm(8) for details.

   Installation from binary rpm:
            % rpm -i gjiten-x.xxx.rpm

   Upgrading from rpm:
            % rpm --upgrade gjiten-x.xxx.rpm

   Creating a debian package from the tarball:
            % tar zxpvf gjiten-x.x.tar.gz
            % cd gjiten-x.x
            % dpkg-buildpackage -rfakeroot
            

   This should create ../gjiten_x.x-1_i386.deb.

   Then you should install this binary package with dpkg -i.
     _________________________________________________________

4. Dictionary files

   Gjiten needs dictionary files to function. You will need the
   kanjidic dictionary file for kanji lookups for KanjiDic and
   other dictionary files for the word dictionary in edict
   format. You can get these from ftp://ftp.cc.monash.edu.au. You
   should download kanjidic.gz and edict.gz. There are some other
   dictionary files also you might be interested in. (j_places,
   compdic, etc.) gunzip the dictionary files to the Path To
   Dictionary Files set in Preferences. This should be the same
   directory where the radkfile and vconj files got installed.
   [default: /usr/share/gjiten/]. The source tarball contains a
   script dicfiles/getdics that will download a bunch of
   dictionary files with wget.

   The Debian distribution contains three dictionary packages
   that you can install via apt-get: enamdict, edict and
   kanjidicic

   As of v2.0 Gjiten uses the GTK2 and GNOME2 libraries. These
   handle text as unicode internally unlike their previous
   versions. Gjiten now needs the dictionary and data files in
   UTF-8 encoding instead of EUC-JP. Please convert the
   dictionary and data files to UTF-8. You can use iconv for
   this.
            % iconv -f EUC-JP -t UTF-8 dictfile -o dictfile.utf8
        

   If the dictionary file is not UTF-8, Gjiten will complain
   after you add them in the preferences.

   The datafiles radkfile.utf8 and vconj.utf8 are needed by
   Gjiten. Please copy these into the directory of your
   dictionary files.
     _________________________________________________________

5. Usage

   Start up gjiten.

   Set the Preferences first, otherwise gjiten might not work
   correctly. Add the dictionary files to the list that you want
   to use with the word dictionary.

   Please note that the kanjidic dictionary file should not be
   added to the list together with the other word dictionary
   files, it has a separate file entry box.
     _________________________________________________________

5.1. Word dictionary usage

   You can set the search options with the toggle buttons.
   Experiment with them a little to see what they do. Select the
   dictionary file in the pulldown menu under general options
   that you want to search in.

   The Auto Adjust Options can be quite a handy feature. Gjiten
   will search for matches setting the English Search Options in
   the following order if it doesn't find any match for your
   prior criteria: Whole Expressions -> Whole Words -> Any
   Matches. Similarly in the case of a Japanese search it will
   set the toggle buttons from top to bottom.

   For Japanese input, you have to activate (shift-space for
   kinput2 in my case) the program on your system that can
   convert and pass Japanese text to GTK. Cut and paste works
   well also! ;)
     _________________________________________________________

5.2. Kanji dictionary usage

   I'll refer to the kanji dictionary part of gjiten as KanjiDic.

   If you know the stroke number of the kanji, enable the Search
   By Strokes option. Set the stroke number. If you are not
   exactly sure, use the +/- field also.

   Enable the Search By Radical option if you know the radical(s)
   of the kanji. Either enter the radical directly through your
   XIM, or click on the Show Radical List for the radical window.
   Here you can click on the appropriate radical to pop it into
   the radical search entry. You can enter up to 10 radicals.

   If you want to search by a keyword also, then enable the
   Search By Key option. A keyword can be anything that the
   Kanjidic file contains. Usually this will be a reading (in
   kana) or an english meaning of the kanji. But it can be a
   kanjidic code also. For example entering S8 here would give
   all the kanji with 8 strokes. Read Jim Breen's documentation
   about the kanjidic file if you want to know more about these.

   KanjiDic has real-time kanji lookup. This means that while you
   are entering radicals it will look up and show the kanji
   matching your criteria without having to press the search
   every time you enter a radical. If only 1 kanji is found then
   an info window will pop up with the information about the
   kanji. If more than one is found then you have to click on the
   kanji to get the info about it. You can customize these info
   fields shown in the Preferences.
     _________________________________________________________

5.3. Command Line

   Issue the command gjiten --help to get a list of all the
   command line arguments that gjiten understands. Here is the
   important part:
 gjiten options
              -k, --kanjidic              Start up Kanjidic instead of
Word dictionary
              -w, --word-lookup=WORD      Look up WORD in first diction
ary
              -l, --kanji-lookup=KANJI    Look up KANJI in kanji dictio
nary
              -c, --clip-kanji            Look up kanji from clipboard
              -v, --clip-word             Look up word from clipboard


   Tip

   As a useful utilization of these I assigned a shortcut
   ctrl-alt-w with my window manager to the command gjiten -v. In
   the Sawfish window manager you can add this command to your
   existing shortcuts as run-shell-command. So I just have to
   highlight the text then press ctrl-alt-w and gjiten gets fired
   up with the search results already looked up. You can add
   another shortcut for the kanji lookup.
     _________________________________________________________

5.4. KanjiPad

   KanjiPad is a separate application written by Owen Taylor. It
   is a handwriting recognition program for kanji. Does a quite
   good job in many cases and might be faster to use then
   KanjiDic. Select the kanji that appears on the right side of
   KanjiPad after recognition, then you can paste this into
   KanjiDic or the into word dictionary for further lookups. Read
   it's README for more info.

   The latest version of KanjiPad can be downloaded from
   http://www.gtk.org/~otaylor/kanjipad/
     _________________________________________________________

6. Troubleshooting

   Here are some common errors and their solutions users usually
   get when running gjiten.
     _________________________________________________________

6.1. File errors

   If you get a message unable to open file: vconj.utf8 or unable
   to open file: radkfile.utf8 then you need to put these two
   data files in the same directory with your dictionary files
   set under the Path To Dictionary Files under Preferences.
   Please read the installation instructions again.
     _________________________________________________________

6.2. Japanese input (mostly Kinput2)

   Gjiten doesn't have any code that has to do with Japanese
   input. The GTK library handles this completely. So if your XIM
   doesn't work, don't blame me! ;)

   Few things that might help to make it work for you: Make sure
   your XIM works fine with other applications. Try Kterm for
   example. If it works with Kterm, then try a GTK+ program.

   Kinput2 needs LC_ALL to be set to ja_JP in order to work,
   otherwise the input window won't pop up. You don't need to
   export this, Gjiten will do it for you if you enable it in the
   Preferences (under fonts).

   You also need the Japanese locale for glibc to be able to
   input with kinput2. Under debian, you will have to put
   ja_JP.EUC-JP EUC-JP into /etc/locale.gen, then rerun
   locale-gen Other distributions should ship this with a locales
   package.
     _________________________________________________________

6.3. Window manager and Kinput2 problems

   There was no solution to this problem in the docs, though
   lately it appeared in the kinput2(1) manual.

   Contributed by John Seebach <jseebach(at)mindless.com>

   Certain window managers (I've seen this problem with
   WindowMaker, Sawmill, and KWM, but there are probably more as
   well) don't handle kinput2 pop-up windows particularly well.
   You'll know you're using one of those Window Managers if you
   fire up gjiten, try to start the conversion widget, and get a
   window that blinks uncontrollably and doesn't allow you to
   enter anything to be converted.

   Fortunately, there are some workarounds. First, try adding the
   following lines to the kinput2 section of your .Xdefaults or
   .Xresources:
  Kinput2*useOverrideShellForMode: true
  *preeditType: OverTheSpot


   Next, you need to tell your window manager how to deal with a
   kinput2 window. I've only tested this with WindowMaker,
   because that's what I use, so outside of that, your on your
   own.

   If you're using WindowMaker, try adding the following lines to
   ~/GNUstep/Defaults/WMWindowAttributes :
  Kinput2 = {
   DontSaveSession = Yes;
   KeepInsideScreen = Yes;
   KeepOnTop = Yes;
   NoAppIcon = Yes;
   NoHideOthers = Yes;
   NoKeyBindings = Yes;
   NoMouseBindings = Yes;
   NoResizebar = Yes;
   NotClosable = Yes;
   NotMiniaturizable = Yes;
   Omnipresent = Yes;
   SkipWindowList = Yes;
   Unfocusable = Yes;
 };


   You may either need to log out of your X session and start
   again, or try:
% xrdb -load .Xresources

   in order for the changes to your X resources to take effect.
     _________________________________________________________

6.4. Fonts and mojibake

   Mojibake occurs when you get garbage instead of Japanese
   characters. With GTK2 and Gjiten, this will be in the form of
   empty boxes in place of Japanese characters. This is a font
   problem.

   Antialiasing (with GDK) and Japanese don't work well at the
   time. You should disable it in the Preferences.

   You also need to select a fontset that has Japanese characters
   to be used with Gjiten. "Sans" and "Serif" and also "unifont"
   should usually work.

   kinput2 needs LC_ALL to be set to ja_JP. If you don't have the
   LANGUAGE environment variable set, then this will also default
   to ja_JP and all the widget text and menus will be in
   Japanese. If you wan't English then check the "Force
   LANGUAGE=C" option in the Preferences.

   The last three options under the Fonts tab need gjiten to be
   restarted for these to take effect, because they are setting
   environment variables.
     _________________________________________________________

6.5. Pasting from Netscape

   If pasting Japanese text from Netscape into Gjiten has no
   effect, only non-japanese can be pasted, then set the LC_ALL
   environment variable to ja_JP in your shell before starting
   Netscape (export LC_ALL=ja_JP for bash). This solved the
   problem in my case.
     _________________________________________________________

6.6. Settings

   Gjiten uses gconf to store its settings, so you can use the
   gconf-editor to poke the settings directly under apps/gjiten.
     _________________________________________________________

6.7. Bugs

   Check for known bugs here first:
   http://gjiten.sourceforge.net/BUGS. If it's not listed here
   and you can reproduce it then report it to me please.
     _________________________________________________________

7. Future Improvements

   Check out http://gjiten.sourceforge.net for the latest
   release. Read the TODO file to see what's planned for future
   releases. Feel free to email me with other suggestions.

   A few people have asked for a --disable-gnome configure option
   (that is a GTK+ only build). IMHO: if you can afford to have
   the GTK+ libs installed on your system, then you can afford
   the Gnome libs also. The gnome libraries provide features that
   would take me much longer to code in pure GTK+. I would have
   to rewrite and add lots of stuff to the sources for this, and
   I just don't have the time. Instead I'd like to add new
   features. If you want to do this, then feel free to hack code
   and send a patch to me.
     _________________________________________________________

8. Feedback

   Send me suggestions, wishes, comments, patches, bug reports,
   money [optional] ;) You can contact me at:
   <boti@rocketmail.com>

   If you send me bugreports, please include the following
   information also:

     * gjiten version
     * distribution and other info about your OS
     * Xfree release
     * gtk release
     * gnome releas
     * output of gdb gjiten and strace gjiten
     * and any other info you might think is important.
     _________________________________________________________

9. Credits

   Thanx go to:

     * Mike Fabian <mfabian at suse dot de> for various patches
       and suggestions,
     * Yamagata Hiroo <hiyori13 at alum dot mit dot edu> for
       translation revisions,
     * John Seebach <jseebach at mindless dot com> for doc and
       makefile fixes,
     * Jason Vertrees <tree at computer dot org> and his
       professor Manabu Mizobe for the Japanese translation of
       the docs and HP, I was to lazy to do it :-)
     * Ryan Nielsen <ran at gondolin dot fortyoz dot org> for his
       feature enhancing patch.
     * Jim Breen <j.breen at csse dot monash dot edu dot au> for
       creating and maintaining the Japanese-English dictionary
       files and xjdic,
     * The Glade people, for making my job easier,
     * People who write free software. You know who you are!