Markus Gunske von Kölln
Co-ordinator Subnet Deu-Nih of Intnl. E-Mail Tandem Network
422 Shizuoka-shi, Shikiji 1-8-6, Sakka Manshon 205, Japan
Tel: (+81)054-238-0999
email markusvk@cc.matsuyama-u.ac.jp
WWW http://cc.matsuyama-u.ac.jp/~markusvk
Kanji in email
or How to set things up
The information in this page and the collected software mentioned below is especially for those people, teachers, system-administrators and users with own systems, who are involved in setting up hardware and software, so that we can work with kanji-email in the Subnet "Deu-Nih" of the International Email Tandem Network.
The collected software is available at two locations. Both locations have the software in the same directory structure. Users in Japan
should download from:
ftp://ftp.cc.matsuyama-u.ac.jp/pub/nih/.
This link has the shortcut "MA".
Users from Europa should download from:
ftp://ftp.slf.ruhr-uni-bochum.de/email/codes/nih/. This link has the shortcut
"BO".
Users from the rest of the world ;-) should use the faster one of the above mentioned links.
Throughout the text I give you the name of the appropriate subdirectory together with the two shortcuts. The shortcuts then point to this
subdirectory (eg. You can find the program xyz.exe in the directory pc/win
(BO/
MA)).
Data on a computer consists of "bits". A bit either has a status of 0 (zero) or 1. The bits are arranged in groups. The most usual group is a "byte" which consists of 8 bit. Due to mathematics a byte can have 256 states (2^8). With this 256 possibilities of a byte the usual character set (ISO 646-1991 and ISO 8859) can be displayed. It not only consists of alphanumericals (0-9, a-z, A-Z) but also of special sequences and special characters, e.g. the umlaute (because of the lack of the umlaute in the forum Deu-Nih (see below) we display them as "o, "a, "u, "s, "O, "A and "U).
But many languages (like the Japanese one) not only consists of 256 symbols but much more. Because of this, the characters are not handled with 1 byte but with 2 byte (16 bit) offering space for up to 65000 symbols. There are different ways of how this number of characters is arranged and how these characters are adressed or encoded.
For the Japanese language there is one main electronic character set standard, JIS X 0208-1990 (with 6879 characters including 6355 kanji) and a supplemental set JIS X 0212-1990 (with 6067 characters including 5801 kanji). For the first there are three encoding methods available: JIS, EUC and Shift-JIS (s. below), for the latter two: JIS and EUC. For JIS X 0208-1990 there are also two preceding sets: JIS C 6226-1978 and JIS X 0208-1983 (which was renamed from JIS C 6226-1983).
These character sets can be encoded on computers in different ways. The best known and most common used are Shift-JIS (also SJIS), EUC (ISO 2022-1993) and JIS (ISO 2022). While EUC is used mainly on unix-machines, Shift-JIS can mainly be found on Macintosh, PC (Dos/J, Win/V, Windows-J) and Japanese PC (NEC PC98xx, Fujitsu FM Towns).
While most computers internally work with 8 bit, the communication on the internet with email is usually done with 7 bit. The 8th bit is simply cut off. More and more hosts change their mail programs to also use and work with 8 bit, but this is not yet common.
The disadvantage of this is, that the kanji and also the german umlaute have to be send as 7 bit encoded data. For kanji there is an appropriate encoding method, JIS, but for umlaute this is not possible. They are in the upper half of the 8 bit standard character set which is covered by the 8th bit and thus cut off.
The JIS encoding works with only 7 bit. This leaves space for 16384 characters, including kanji, kana, romaji, numerals, special characters and sequences. The characters in the Japanese standard character sets are accessed with so called escape sequences. For each standard character set there are different sets of escape sequences. The escape sequence set for JIS C 6226-1978 is called Old-JIS the one for JIS X 0208-1983 is called New-JIS. Many people address the set for JIS X 0208-1990 also with New-JIS.
So, what we need is software which works internally with whatever coding it things is OK but externally only sends 7 bit JIS coded data onto the net.
Unix-Systems most often use the EUC encoding internally. With a mail program you have to filter the mail and convert it to JIS.
Fred Bremmer at
Matsuyama University (Japan) has written a pine-patch which handles the conversion
automatically. The host running at Matsuyama University runs SunOS Release 4.1.3-J
LE1.1.3_U1. Together with Pine V3.90 and this patch we are cabable of doing Japanese email with 7 bit JIS. More information
on the patch you can find at ftp://ftp.cc.matsuyama-u.ac.jp/pub/pine/ and in the
directory unix/ (BO /
MA) you can find the Pine 3.90 sources, so you can compile your
version of this patched pine. For SunOS with the "Japanese Language Extension" it is important that the KCODE enviroment
variable is set to EUC (e.g. setenv KCODE EUC). Please make sure that setterm is NOT set to SJIS. Now
everybody working on the host console can do Japanese email.
The equivalent should also work on other unix systems with Japanese language support and the modified pine (though not yet testet).
Please remember to compile everything!
At Matsuyama University the students work mostly with PC 98xx. They (the PC ;-) are connected with a lan to the host. On this side work can be done with a modified NCSA Telnet. The NCSA Telnet and the language modification can be found at in the directory pc98xx/dos/ (BO / MA). With this telnet we log onto the host and work with a modified pine (see Chapter 2) in our accounts. For this, unix must have been set up as described above.
Another way is working with Windows on this machines. For this, a Japanese mail program must be installed on the computer. Examples of this programs you can find in the directory pc98xx/win/ (BO / MA).
These PC usually have software like Dos/V, Win/V or Windows-J. The procedure is similar to that of the Japanese computers (NEC, Fujitsu). For information about how to set things up, please look in chapter 3 (NEC PC 98xx / Fujitsu FM Towns).
Software which can be used on this computers can be found in directory pc/winvj/ (BO / MA) for Windows (/V, -J) and in directory pc/dosvj/ (BO / MA) for Dos/V.
This section is divided into two parts, one for the DOS-World and one for the Windows-World (also Windows95). The main procedures for both are almost identical. Working with the PC offers two choices, either you work with a telnet on a Unix-Host and you are mailing there or you get your email on your local system with a dedicated mail program and mail there. Windows has one great advantage. There are programs with which you can view Japanese texts (E-Mail, News, WWW) even if you don't have any other support for Japanese languages on your computer. The disadvantage of these programs is that you can't enter or edit text in Japanese. For more about this see the practical part below.
Working with telnet means that you have to save your Japanese mail in your unix-mailer (e.g. elm, pine) to a file, then download this
file with ftp to your local PC. Here you can read this file with a Japanese word-processor. Writing the message is the same procedure
backward. You have to write the message with the word-processor and save it to a file.
Now the important part: You must save this file in JIS format!!
After this you can transfer the file with ftp to your remote Unix-Host and load it in your mail program. You will only see garbage on the
screen, but this is what it looks like in JIS format without language support. Your mail is now ready to be sent.
Working with a local mail program is more simple. You receive the Japanese mail and looking at it will reveal you lots of garbage (if you
have not enabled the Windows Japanese language view facility). In your mail program save this mail to a file and load this file in your
Japanese word-processor. Now it should look like kanji. Sending mail is vice versa. Write the mail with your word-processor, save it to a file
(don't forget to use JIS format!) and then load it into your mailer. Now the message is ready to go.
Due to the fact that most PC software working with kanji uses Shift-JIS coding internally, don't be confused if your word-processor converts
the incoming JIS coded file to Shift-JIS. This is only for its internal work. Save your new file in JIS format!
The practical part looks like this:
If you work with a telnet, you can use NCSA Telnet without the language support, files can be found in the directory pc/dos/ (BO / MA). In this package an ftp-client is also available. The shareware Japanese word-processor NJStar 3.0 for DOS is also available in the directory pc/dos/ (BO / MA). A dedicated mail program for Dos is "Pegasus Mail for Dos".
A telnet and ftp client is usually in the software package which connects you to Internet. If you want some more and probably better choices, have a look at Strout's Consummate Winsock Apps List. There you can also find mail clients. The, in my humble opinion, best mail client "Pegasus Mail for Windows V2.x" can also be found in the directory pc/win/ (BO / MA). The probably best Japanese word-processor for windows "JWP" can also be found in the directory pc/win/ (BO / MA).
In Windows you can also work with "Cut/Copy and Paste". If you do this, and your word-processor gives you the opportunity, please don't forget to change the clipboard-format to JIS. In JWP this can be done with the "Utilities|Options" menu entry. In JWP you can choose between Old-JIS and New-JIS, please use New-JIS (see also information about encoding systems in chapter 1).
For Windows there are programs which enable you to view Japanese texts (Shift-JIS, JIS and EUC) in WWW pages, Usenet news and e-mail. With this programs you can preview the incoming Japanese mail and only when there is an important mail or one you would like to answer to then you have to use the above described methods to write a new message (or reply to one). This programs can be found in the directory pc/win/ (BO / MA).
On the Macintosh at Matsuyama University the English System 7.5.1 with Japanese Language Kit is running. The users of this machines either connect with a telnet program or with Mac Eudora 1.3.7-J12, the Japanese version of this freeware program, to the internet. With telnet they work with the modified pine (see also the information about Unix), with Eudora they directly work on Japanese email and Eudora automatically converts it to JIS code. In Germany equivalent systems are running. The difference is, that they use German Systems 7.5.1 or 7.1 but also with Japanese Language Kit and mostly Eudora-J software. I provide a copy of Mac Eudora-J and NCSA Telnet in the directoy mac/ (BO / MA).
Here I would like to give you some more resources of where to find valuable information on processing the Japanese language on computers. Some of this resources I used to compile this text.
All the software mentioned above can also be obtained at different sources. A good archive to start searching is always the Monash University Nihongo Archive (ftp://ftp.cc.monash.edu.au/pub/nihongo) and the other in chapter 7 mentioned resources.
Markus Gunske vonKölln
E-Mail: markusvk@cc.matsuyama-u.ac.jp
422 Shizuoka-shi, Shikiji 1-8-6, Sakka Manshon 205, Japan
Tel: (+81) 054-238-0999
Last Updated: Monday, 24. February 1997 - 04:44 MET
To top of page.