You are not logged in.

#1 2005-02-15 16:12:44

WaFuSan
Member
From: Nagoya
Registered: 2005-01-27
Posts: 39
Website

gjiten - japanese dictionary

In the japanese applications list...

PKGBUILD

# Contributer: David Mezerette <david.mezerette@lapis.plala.or.jp>
pkgname=gjiten
pkgver=2.3
pkgrel=1
pkgdesc="A Japanese dictionary for GNOME"
depends=('xorg' 'libgnomeui')
suggests=('kanjipad')
install=gjiten.install
url="http://gjiten.sourceforge.net/"
license=GPL
source=(http://$pkgname.sourceforge.net/$pkgname-$pkgver.tar.gz)


build() {
  cd $startdir/src/$pkgname-$pkgver
  ./configure --prefix=/usr 
  make || return 1
  make DESTDIR=$startdir/pkg/ install
} 

I have included in the gjiten.install file the script from the gjiten homepage  http://gjiten.sourceforge.net/ to download the EDICT and the like english-japanese dictionaries (all open source too).
gjinten.install

#!/bin/sh

# arg 1:  the new package version
post_install() {
  #
#!/bin/sh
cd /usr/share/gjiten

# BEGINNING OF THE DICTIONARIES DOWNLOAD SCRIPT
# FROM http://gjiten.sourceforge.net/

mkdir dics
mkdir dl
mkdir docs
mkdir tmp

#EDICT
wget -c http://ftp.cc.monash.edu.au/pub/nihongo/edict.gz -O dl/edict.gz
gunzip   <dl/edict.gz >tmp/edict
iconv -f EUC-JP -t UTF-8 tmp/edict -o dics/edict

#Japanese Names
wget -c http://ftp.cc.monash.edu.au/pub/nihongo/enamdict.gz -O dl/enamdict.gz
gunzip   <dl/enamdict.gz >tmp/enamdict
iconv -f EUC-JP -t UTF-8 tmp/enamdict -o dics/enamdict

#Kanjidic
wget -c http://ftp.cc.monash.edu.au/pub/nihongo/kanjidic.gz -O dl/kanjidic.gz
gunzip   <dl/kanjidic.gz >tmp/kanjidic
iconv -f EUC-JP -t UTF-8 tmp/kanjidic -o dics/kanjidic

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/jddict.v02.gz -O dl/jddict.gz
gunzip   <dl/jddict.gz >tmp/jddict
iconv -f EUC-JP -t UTF-8 tmp/jddict -o dics/jddict

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/lifscdic.Z -O dl/lifscdic.gz
gunzip   <dl/lifscdic.gz >tmp/lifscdic
iconv -f EUC-JP -t UTF-8 tmp/lifscdic -o dics/lifscdic

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/compdic.gz -O dl/compdic.gz
gunzip   <dl/compdic.gz >tmp/compdic
iconv -f EUC-JP -t UTF-8 tmp/compdic -o dics/compdic

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/engscidic.gz -O dl/engscidic.gz
gunzip   <dl/engscidic.gz >tmp/engscidic
iconv -f EUC-JP -t UTF-8 tmp/engscidic -o dics/engscidic

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/j_places.gz -O dl/j_places.gz
gunzip   <dl/j_places.gz >tmp/j_places
iconv -f EUC-JP -t UTF-8 tmp/j_places -o dics/j_places

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/geodic.gz -O dl/geodic.gz
gunzip   <dl/geodic.gz >tmp/geodic
iconv -f EUC-JP -t UTF-8 tmp/geodic -o dics/geodic

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/compverb.gz -O dl/compverb.gz
gunzip   <dl/compverb.gz >tmp/compverb
iconv -f EUC-JP -t UTF-8 tmp/compverb -o dics/compverb

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/4jwords.gz -O dl/4jwords.gz
gunzip   <dl/4jwords.gz >tmp/4jwords
iconv -f EUC-JP -t UTF-8 tmp/4jwords -o dics/4jwords

#########Uncompressed######
#this doesn't work well, because of it's bad format.
#wget -c http://ftp.cc.monash.edu.au/pub/nihongo/pandpgls.euc -O dl/pandpgls
#iconv -f EUC-JP -t UTF-8 dl/pandpgls -o dics/pandpgls

##BAD FORMAT
#wget -c http://ftp.cc.monash.edu.au/pub/nihongo/lawgloss.euc -O dl/lawgloss
#iconv -f EUC-JP -t UTF-8 dl/lawgloss -o dics/lawgloss

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/jisterme.wip -O dl/jisterme
iconv -f EUC-JP -t UTF-8 dl/jisterme -o dics/jisterme


#ZIP
wget -c http://ftp.cc.monash.edu.au/pub/nihongo/classical.zip -O dl/classical.zip
unzip -o  dl/classical.zip -d tmp
iconv -f EUC-JP -t UTF-8 tmp/classical -o dics/classical
mv tmp/classical.euc docs/classical.txt

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/lingdic.zip -O dl/lingdic.zip
unzip -o  dl/lingdic.zip -d tmp
iconv -f EUC-JP -t UTF-8 tmp/lingdic -o dics/lingdic
mv tmp/lingdic.txt docs/lingdic.txt

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/forsdic.zip -O dl/forsdic.zip
unzip -o  dl/forsdic.zip -d tmp
iconv -f EUC-JP -t UTF-8 tmp/forsdic_e -o dics/forsdic
mv tmp/forsdic_e_doc.txt docs/forsdic.txt

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/aviation.zip -O dl/aviation.zip
unzip -o  dl/aviation.zip -d tmp
iconv -f EUC-JP -t UTF-8 tmp/aviation -o dics/aviation
mv tmp/aviation.txt docs/aviation.txt

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/concrete.zip -O dl/concrete.zip
unzip -o  dl/concrete.zip -d tmp
iconv -f EUC-JP -t UTF-8 tmp/concrete -o dics/concrete
mv tmp/concrete.doc docs/concrete.txt

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/findic.zip -O dl/findic.zip
unzip -o  dl/findic.zip -d tmp
iconv -f EUC-JP -t UTF-8 tmp/findic -o dics/findic
mv tmp/findic.doc docs/findic.txt

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/mktdic.zip -O dl/mktdic.zip
unzip -o  dl/mktdic.zip -d tmp
iconv -f EUC-JP -t UTF-8 tmp/mktdic -o dics/mktdic
mv tmp/mktdic.doc docs/mktdic.txt

wget -c http://ftp.cc.monash.edu.au/pub/nihongo/stardict.zip -O dl/stardict.zip
unzip -o  dl/stardict.zip -d tmp
iconv -f EUC-JP -t UTF-8 tmp/stardict -o dics/stardict

rm -rf tmp

echo ""
echo ""
echo "---------------------------"
echo UTF-8 Dicfiles are in /usr/share/gjiten/dics/
echo "---------------------------"
# END OF THE DICTIONARIES DOWNLOAD SCRIPT

echo ""
echo "gjiten: don't forget to add an input method for japanese"

  /bin/true
}

# arg 1:  the old package version
post_remove() {
  #
  # remove the dictionary files and the gjiten directory
  rm -r /usr/share/gjiten
  #
  /bin/true
}

op=$1
shift
$op $*

next on the list: kanjipad, to "draw" kanji (and which can be launched from gjiten)

Offline

#2 2005-02-15 16:24:25

scottro
Member
From: NYC
Registered: 2002-10-11
Posts: 466
Website

Re: gjiten - japanese dictionary

Actually, it'd be nice if you put together a canna and kinput2 canna build.  I tried but couldn't get one to work, though I do have a vanilla tarball with some modifications that works (not as an Arch package, but with the usual
make make install stuff)


Whiningly yours

Offline

#3 2005-02-15 16:33:10

WaFuSan
Member
From: Nagoya
Registered: 2005-01-27
Posts: 39
Website

Re: gjiten - japanese dictionary

Well, I posted some very unsatisfying (i.e. not arch-compliant) anthy, scim-anthy (for japanese) and scim-hangul (for korean) PKGBUILDs yesterday... which is my favorite input method.
The point of not including them is not to impose an input method. I have tried a few, and I guess it's really a matter of personal preferences.
I will check the PKGBUILD of those and improve them...

Offline

#4 2005-02-21 09:01:01

neri
Forum Fellow
From: Victoria, Canada
Registered: 2003-05-04
Posts: 553

Re: gjiten - japanese dictionary

could it be that this post and the answers has been vanished? In case it is lost I will repost it...
-neri

Offline

#5 2005-02-21 13:05:40

WaFuSan
Member
From: Nagoya
Registered: 2005-01-27
Posts: 39
Website

Re: gjiten - japanese dictionary

It looks like they are lost indeed.
I read the post this morning, thought I will answer it a bit later, and it disappeared meanwhile.

Offline

#6 2005-02-21 15:14:10

neri
Forum Fellow
From: Victoria, Canada
Registered: 2003-05-04
Posts: 553

Re: gjiten - japanese dictionary

WaFuSan wrote:

It looks like they are lost indeed.
I read the post this morning, thought I will answer it a bit later, and it disappeared meanwhile.

I talked to rasat, he told me thatsome little problems occured, so I will
repost it, hope that I remeber most of it...

It's about to take som eaction on i18n in Arch in general. The devs got a
mail by Charly, who basically wanted to negotiate between th CLE people
and Arch-Devs since he speaks either Chinese and English. He described
things here bit: http://bbs.archlinux.org/viewtopic.php?t=10079

Now I replied, here are the basic points:

None of the Arch-developer comes from an area where support for big
alphabets like Chinese-trad/modern or Japanese is needed. Support for
such stuff was only added on request (xpdf font mappings). Now I see it
is time to have this a bit more organized and there probably will be
such an i18n department in Arch's developer structure. And maybe it's
me, who will take action on that. I like to make it possible to support
languages and input methods in the most generic way, if this is
possible. Just to make sure most languages _can_ be supported and
things for one language doesn't block things for others. The whole
computer industry still suffers from the fact that computers have
been developed in English speaking countries. I've made some research on
that, but you must know that I don't speak any of these languages. From
what I found, Arch needs the following things:

- good input methods (from my readings uim and scim look promising, I'm
  favouring scim as it is supported by freedesktop.org and it seems
  to can or will be able in future things like XIM or IIIMF)

- Conversion engines like anthy (which is kanna-kanji AFAIR)

- m17n library

- fonts, supporting most glyphs, readable stuff, not "beautiful"

Unfortunately, from the CLE website I can not gather what they are
using, providing and which underlaying framework is used. I don't even
see which fields of i18n are covered.

maybe you can describe things a bit, too me, eventually, it will help
both of us

See, my point is I don't wanna prefer any language like Chinese and
leave Japanese an Korean out.

To sum up:
- is SCIM a way to go, how are user expieriences with it?
- what else is needed
- where do we get fonts and which fonts are needed
- what does the CLE patches provide, do they help us also by providing a framework for other languages?

I prolly forgot something from the previos post, but anyway thes are my questions smile

Thanks fro reading,
-neri

Offline

#7 2005-02-21 16:14:23

WaFuSan
Member
From: Nagoya
Registered: 2005-01-27
Posts: 39
Website

Re: gjiten - japanese dictionary

First, the few infos google gave me about CLE

CLE stands for Chinese GNU/Linux Extensions, it is a collection of Chinese related software on GNU/Linux platform. CLE is based on and optimised for the Red Hat Linux distribution. As with Red Hat, software is packaged in the RPM format. CLE also support other major distributions such as Slackware and Mandrake. You will get a complete Chinese (BIG5/GB) environment under Linux by installing CLE.

from http://distrowatch.serve-you.net/table. … bution=cle

One of the easiest way of getting Chinese support in Slackware is to install CLE (Chinese Linux Extension), it's very famous in Taiwan.

Homepage: (If you can read Chinese)
http://cle.linux.org.tw/

Slackware packages:
http://cle.linux.org.tw/Slackware/

P.S. You don't need to install all packages if you just want to view/input Chinese, install all packages will localize your Slackware. You can read install-en.txt for more information.

from http://www.linuxquestions.org/questions/history/282399

So, from what I got, CLE provides packages, ranging from the fonts to a full chinese environment. I don't know the details though, but I would be interested into knowing what input method they use.

To answer the questions:
In my opinion, SCIM is a very good input method: i use it for japanese and korean, and it supports an incredibly high number of languages (many of them I have never heard of, btw. Haven't found klingon though  wink  ).
To make it run, you need:

makedepends=('intltool')
depends=('xorg' 'gtk2' 'gconf') 

(from http://bbs.archlinux.org/viewtopic.php?t=7920 plus the different input method (scim is just the platform):
- scim-anthy
- scim-hangul
- scim-pinyin
- scim-m17n...
(cf http://www.scim-im.org/projects/imengines for details)

As for the fonts, from what I know, the most general ones are:
- the MS Mincho series
- the Bitstream Cyberbit
Then japanese users will probably request more fonts, chinese users too, korean users too... if you begin going into fonts, it looks like a jungle (and there are conflicts over how some characters, shared by japanese and chinese for instance but slightlt different, should be written)

Now, my opinion:
Two things should be kept separate:
- the input method
- the localization
I appreciate to be able to input non-roman languages (I have to in fact^^), but to keep an english desktop. To do this, an input method is enough. If scim were adopted, we can imagine a scim "meta-package", asking at install which IM should be installed at the same time (like the pacman -Sy kde which asks for installation of kde-tools, kde-base and so on...)
The localization would be appreciated by native users, and it probably means a package for every language (like the debian apt-get install locale-ja I guess... not a debian specialist, but I tried it on a Mepis and got a full japanese environment after setting the locales)

Now, I read enough japanese to understand the install files, if that can help...  big_smile

Ah, almost forgot: another thing that our chinese/korean/japanese (see, I alternate the order^^) would also appreciate is the CJK extension to LaTeX.

Offline

#8 2005-02-22 00:01:01

neri
Forum Fellow
From: Victoria, Canada
Registered: 2003-05-04
Posts: 553

Re: gjiten - japanese dictionary

WaFuSan wrote:

So, from what I got, CLE provides packages, ranging from the fonts to a full chinese environment. I don't know the details though, but I would be interested into knowing what input method they use.

well, we don't even have localized western versions of some packages, and ship only what comes with the sources. To achieve a fully Chinese environment would be a bit much smile, but the fonts sounds interesting, and for the imput method we shold try to focus on SCIM, other can be added later hwen we have one thing running stable.

WaFuSan wrote:

To answer the questions:
In my opinion, SCIM is a very good input method: i use it for japanese and korean, and it supports an incredibly high number of languages (many of them I have never heard of, btw. Haven't found klingon though  wink  ).
To make it run, you need:

makedepends=('intltool')
depends=('xorg' 'gtk2' 'gconf') 

(from http://bbs.archlinux.org/viewtopic.php?t=7920 plus the different input method (scim is just the platform):
- scim-anthy
- scim-hangul
- scim-pinyin
- scim-m17n...
(cf http://www.scim-im.org/projects/imengines for details)

these are some good links and stuff. Main focus should be to have a well running scim-base package, hoping that the actual im can be build easier. How is that scim started in X btw?
From my readings i figured that must be in .xinitrc/.xsession:

export GDK_USE_XFT=1
scim -d
export GTK_IM_MODULE="scim"
export XMODIFIERS="@im=SCIM"

I just wonder if that scim -d means we could build it as deamon with an rc.d/scim script. And moreover, what is needed to make it working with qt/kde-apps or is that gtk2 stuff only needed to povide the Gui for scim?

WaFuSan wrote:

As for the fonts, from what I know, the most general ones are:
- the MS Mincho series
- the Bitstream Cyberbit
Then japanese users will probably request more fonts, chinese users too, korean users too... if you begin going into fonts, it looks like a jungle (and there are conflicts over how some characters, shared by japanese and chinese for instance but slightlt different, should be written)

These looks not too good sad MS Mincho is probably not free, and legally only available to users that own copy of winXP. And for the Bitstream fonts only the Very family is freely available. So this will turn out to be tricky.

WaFuSan wrote:

Now, my opinion:
Two things should be kept separate:
- the input method
- the localization
I appreciate to be able to input non-roman languages (I have to in fact^^), but to keep an english desktop. To do this, an input method is enough. If scim were adopted, we can imagine a scim "meta-package", asking at install which IM should be installed at the same time (like the pacman -Sy kde which asks for installation of kde-tools, kde-base and so on...)
The localization would be appreciated by native users, and it probably means a package for every language (like the debian apt-get install locale-ja I guess... not a debian specialist, but I tried it on a Mepis and got a full japanese environment after setting the locales)

right, and foremost we should focus the input/display capabilities, localization is a big field that can never be fully covered I fear.

WaFuSan wrote:

Ah, almost forgot: another thing that our chinese/korean/japanese (see, I alternate the order^^) would also appreciate is the CJK extension to LaTeX.

I will ask Xentac about this, he maintains tetex.

these explianantions have been helpful for me,
Let me sum up, maybe we can make a roadmap or something:
- scim  (if possible as a deamon)
- anthy (which probably is a dependency for scim-anthy)
- scim-anthy
- scim-chinese (I hope ganlu can help here)
I wonder about these, ganlu packaged scim-chinese but sources seem to have disappeared, but scim-im.org list scim-chewing(and it's dep libchewing) and scim-pinyin as im for Chinese, what is right/better?
- scim-hangul

- scim-tables (who can test these ?)
- m17n and scim-m17n (what exactly does that ?!?)

One thing is still open, where do we good, readable, printable and free fonts?

Also another thing (yes i might be stupid smile ): does anthy really does the same thing as the canna server or is there something more needed by the Japanese writers?

-neri

Offline

#9 2005-02-22 02:40:12

WaFuSan
Member
From: Nagoya
Registered: 2005-01-27
Posts: 39
Website

Re: gjiten - japanese dictionary

How is that scim started in X btw?
From my readings i figured that must be in .xinitrc/.xsession:
Code:
export GDK_USE_XFT=1
scim -d
export GTK_IM_MODULE="scim"
export XMODIFIERS="@im=SCIM"

I just wonder if that scim -d means we could build it as deamon with an rc.d/scim script. And moreover, what is needed to make it working with qt/kde-apps or is that gtk2 stuff only needed to povide the Gui for scim?

Well, when I installed it, even without this script, scim was started with X. I thought it wouldn't, but.... The problem is I could use it with mozilla and gaim, but not with openoffice or kwrite/kedit/kxxx. I needed that script to achieve that.
There is a KDE frontend for scim, called skim, which is probably not necessary to run qt/kde-apps. I will make a test a little later and tell you the results. When using skim, the "scim -d" line is not necessary anymore.
To quote the skim-hanbook FAQ http://scim.sourceforge.net/skim/doc/us … troduction

4.1.
Question: Why do you bother to write skim now that scim-lib already has a Gtk2 based panel?
Ok, the most important reason is that: Nearly all applications running on my Desktop are KDE/Qt based ones, and scim-lib's panel was the only exception, and it does not fit into the entire K Desktop Environment very well, so I decided to write a KDE/Qt panel for scim-lib.

scim runs as a daemon, and is activated by the CTRL+space key combination (which can be redefined. I personnaly use CTRL+ALT+space, because CRTL+space is used for unsecable spaces in open office)


These looks not too good Sad MS Mincho is probably not free, and legally only available to users that own copy of winXP. And for the Bitstream fonts only the Very family is freely available. So this will turn out to be tricky.

Fonts will be a problem indeed...
Though a single font would be IMO the simplest, it is maybe not the best:

Covering Unicode completely with a single font is usually neither necessary nor desirable. It is often better to use several fonts at once to cover the subset of Unicode you need because it often results in typographical more pleasing results. For example, in most CJK-fonts the Latin glyphs are less well designed than the glyphs in fonts specialising on European languages. And, because of the Han-Unification 7 a single font covering all of Unicode cannot satisfy the style preferences of all CJK users. Therefore it is often better to combine a nice font for Latin glyphs with a nice font for Chinese etc..

(from http://www.suse.de/~mfabian/suse-cjk/fo … icode.html)
The important thing is: usually not necessary which means 1- that the let's say arabic user wanting to input chinese will have some trouble 2- that the user has a minimum of knowledge about what fonts to use. A good "read" could be http://www.alanwood.net/unicode/fontsbyrange.html

For example, on suse, one has the following packages for true-type fonts:
package name / language
- ttf-kochi-mincho / Japanese     
- ttf-kochi-gothic / Japanese     
- kochi-substitute / Japanese     
- baekmuk-ttf  / Korean     
- ttf-arphic-gbsn00lp  / simplified Chinese     
- ttf-arphic-gkai00mp / simplified Chinese     
- ttf-arphic-bkai00mp  / traditional Chinese     
- ttf-arphic-bsmi00lp / traditional Chinese
(still according to http://www.suse.de/~mfabian/suse-cjk/av … fonts.html)

- scim-tables (who can test these ?)
- m17n and scim-m17n (what exactly does that ?!?)

I can make the packages as soon as I get back home, it is probably just a ./configure make make install anyway ^^
scim-tables (I have it installed on a Mepis) proposes input methods for japanese (but it is not really usable since it comes with no "word" definition, you have to enter the words kanji after kanji), korean (but my gf, native korean, found it somewhat strange - she's used to MS IME) , chinese and cyrillic (about which I know nothing)
m17n proposes the following IM:
- Amharic
-  Arabic (Egypt)
- Assamese
- Bengali
- Tibetan
- Greek
- Persian
- Gujarati
- Hebrew
- Indi
- Croatian
- Kazakh
- Cambodian
- Kannada
- Japanese
- Korean
- Laothian
- Malayalam
- Oriya
- Panjabi
- Russian
- Slovak
- Serbian
- Tamil
- Telugu
- Thai
- Vietnamese
- Chinese Simplified

As you can see, some overlap, but I guess they differ in some way.

The reason why my preference goes to anthy/scim-anthy for japanese is that it is easily possible to add new words to the dictionary used for the transcription kana (which is what you type) to the word you want, kind of like the personal dictionary for a spell checker.

Also another thing (yes i might be stupid Smile ): does anthy really does the same thing as the canna server or is there something more needed by the Japanese writers?

As an occasional japanese user (for mails and chat), I have seen no difference between the kanna+kinput2 method and the anthy+scim+scim-anthy (except for the dictionary thingie).
It also seems to be more easy to install and configure (almost don't have to play with the locales ^^)

Offline

#10 2005-02-22 10:19:46

cmp
Member
Registered: 2005-01-03
Posts: 350

Re: gjiten - japanese dictionary

just for curiosity - what's scim-anthy?
I created once scim, and scim-table packages, and the standarad input method for japanese, is currently sufficient for me - I just started learning japanese, so I do mostly input kana.

Offline

#11 2005-02-22 10:29:04

neri
Forum Fellow
From: Victoria, Canada
Registered: 2003-05-04
Posts: 553

Re: gjiten - japanese dictionary

cmp wrote:

just for curiosity - what's scim-anthy?
I created once scim, and scim-table packages, and the standarad input method for japanese, is currently sufficient for me - I just started learning japanese, so I do mostly input kana.

Well, I should better shut up if I don't know things right but it has something
to do with canna->kanji conversion, and does what cannaserver used to do in
other(older) systems: some interseting readings:
http://www.jw-stumpel.nl/stestu.html
http://www.jw-stumpel.nl/input.html
mostly debian focused, though

-neri

Offline

#12 2005-02-22 10:54:46

WaFuSan
Member
From: Nagoya
Registered: 2005-01-27
Posts: 39
Website

Re: gjiten - japanese dictionary

neri wrote:

Well, I should better shut up if I don't know things right but it has something
to do with canna->kanji conversion,

Well, you're perfecly right.
At first, I only installed scim-tables, and if I wanted to enter a word like 動物 (animal), I had first to enter どう, convert it to the 動 kanji, then the same with ぶつ->物.
With anthy, I can enter どうぶつ and it will be converted as a single word 動物.
Besides, I couldn't find in scim-tables the "chiisai tsu" character, like in いっぱい (it is the second one, which is usually obtained by typing twice the same consonne).

By the way cmp, what do you call the standard input method for japanese ??   big_smile

Oh, I have done the PKGBUILD for kanjipad as well:
http://bbs.archlinux.org/viewtopic.php?t=10173

Offline

#13 2005-02-22 11:56:20

cmp
Member
Registered: 2005-01-03
Posts: 350

Re: gjiten - japanese dictionary

with standard input I mean the module called nippon, but you're right it's really annoying to input kanji, but as I said, I'm a very beginner and just now arround 15 kanji. so it's sufficient for now big_smile.
the chiisai tsu, can be obtained by typing _tsu with the table/module HIRAGNA. the same for the small a,e,i,o,u.

by the way, do you now any good vocabulary trainer for japanese? - currently I'm using a simple bash script.

and another question: I can use scim only in gtk programms. In the links posted above it's stated that I have to switch to a japanese locale to activate scim for all x programms.
Is there any other way?

Offline

#14 2005-02-22 23:13:40

WaFuSan
Member
From: Nagoya
Registered: 2005-01-27
Posts: 39
Website

Re: gjiten - japanese dictionary

cmp wrote:

the chiisai tsu, can be obtained by typing _tsu with the table/module HIRAGNA. the same for the small a,e,i,o,u.

Thank you, I didn't know... but it isn't very convenient, is it ? tongue

cmp wrote:

by the way, do you now any good vocabulary trainer for japanese? - currently I'm using a simple bash script.

Well, I know there are a few kanji flashcards programs out there, and I think kanjidrill might be a vocabulary trainer, if you add the "usefiles". I don't know for sure, tried it once, thought "cool, wish I had the time" and forgot about it. Maybe I will try again. Anyway, here is the homepage:
http://www.bolthole.com/kdrill/

cmp wrote:

and another question: I can use scim only in gtk programms. In the links posted above it's stated that I have to switch to a japanese locale to activate scim for all x programms.

Sorry, maybe I wasn't clear. It was necessary with canna+kinput2, not with scim.
Here is the script I have in my .xinitrc file:

export GDK_USE_XFT=1
#scim -d
export XMODIFIERS="@im=SCIM"
export GTK_IM_MODULE="scim"
export QTM_IM_SWITCHER=imsw-multi
export QT_IM_MODULE="scim"
export LANG=en_US.UTF-8

And here are the results of a few tests I just did. gaim was used as a test program for gtk applications, and kwrite for kde's.

- A: skim + above script: skim appears in the system tray, works fine with both gtk and kde applications
- B: skim + commented script: skim appears in the system tray, works fine with gtk applications, not with kde's
- C: no skim + commented script: scim is activated by the activation keys (default:CTRL+space) in gtk applications, but not in kde's.
- D: no skim + above script: same as C
- E: no skim + above script + uncommented line "scim -d": scim appears in the system tray, works fine with both gtk and kde applications.

I have made a PKGBUILD for skim, I will test it some more and then post it in the scim thread. [edit: posted]
Hope that helped  big_smile

Offline

#15 2005-02-23 17:09:56

cmp
Member
Registered: 2005-01-03
Posts: 350

Re: gjiten - japanese dictionary

I'm actually not concerned about kde apps, but simple x apps, like xterm, because sometimes it would be nice, if I could enter japanese.
my .bashrc:

#enable uft 8
export LANG=en_US.UTF-8
export LC_ALL=$LANG
export LESSCHARSET=utf-8

# enable scim
export XMODIFIERS=@im=SCIM
export GTK_IM_MODULE="scim"

and then 'scim -d' is launched in my xinitrc file.

Offline

#16 2005-02-24 00:16:22

scottro
Member
From: NYC
Registered: 2002-10-11
Posts: 466
Website

Re: gjiten - japanese dictionary

The best way I found to develop vocabulary was to have imaginary conversations with people--then, I would realize I didn't know a word and find out how to say it.

Living in NYC, I was fortunate, I did a language exchange thing with people wanting to learn English.  (I also studied at a formal school, but just learned textbook style Japanese there.)

There is a very large Japanese population in NYC and the language exchange pretty much became my social life for several years.   Some weekends, the only time I'd speak English, ironically enough, was at the karate doujou.

Also ironically, after marrying a Japanese woman, my Japanese went down hill. Her English is excellent and she prefers to speak it   Worse, she chased away all the cute  air head girls who couldn't speak Engilish, thus depriving me of some incentive to practice.  smile

Seriously, the imaginary conversation is an EXCELLENT way to improve vocabulary.

Offline

#17 2005-02-24 04:01:28

WaFuSan
Member
From: Nagoya
Registered: 2005-01-27
Posts: 39
Website

Re: gjiten - japanese dictionary

cmp wrote:

I'm actually not concerned about kde apps, but simple x apps, like xterm, because sometimes it would be nice, if I could enter japanese.

With my current condiguration (A in the previous post: skim + the script in the .xinitrc), I can enter japanese in xterm. wink

Offline

#18 2005-02-24 16:23:09

cmp
Member
Registered: 2005-01-03
Posts: 350

Re: gjiten - japanese dictionary

hm, strange, but I do not have kdelibs installed, and would to prefer to keep it that way, anyway thanks for the help so far.

@scottro living in a small town in germany I don't know any japanese people, so I think I have to stick to imaginary japanese people wink.

Offline

#19 2005-02-26 01:51:06

WaFuSan
Member
From: Nagoya
Registered: 2005-01-27
Posts: 39
Website

Re: gjiten - japanese dictionary

with the above "E configuration" (no skim + script iin .xininrtc + uncommented line "scim -d"), I can input japanese in xterm.

Maybe you should check the /etc/scim/global files, and look for the /SupportedUnicodeLocales, which should be set up to your locales.
That's just an idea though,  not sure it will work.

Offline

#20 2005-02-26 10:14:05

cmp
Member
Registered: 2005-01-03
Posts: 350

Re: gjiten - japanese dictionary

thanks, does work, I just had to add an entry for en_US-UTF8:
/DefaultIMEngineFactory/en_US-UTF8 = 61ea7bfd-d62a-4dd6-842a-c7f91db5d6ff

Offline

Board footer

Powered by FluxBB