README By Ross Ridge Public Domain @(#) myjis README 1.4 97/08/04 04:59:52 README for myjis. This package is an BETA release of myjis, a set of library routines for working with Japanese text. See the file version.h for the release number. It also includes a set of utilities built with this library, including one for for looking up words in Jim Breen's Japanese to English dictionary EDICT. These libraries and utilities are still under development are not intended for general use. Configuring, installing, and using this package shouldn't be done by any one not willing to go to some effort to do so. There is no automatic configuration utility, and the install target probably won't work for you. You'll need several things not supplied with this distribution to compile and use the utilities in this library. Installed on your system you'll need a reasonably POSIX compliant shell (eg. ksh or bash), a modern version of awk (eg. nawk or gawk), as well as good ANSI C conforming development system. Not necessary, but potentially handy if your system can't display Japanese characters, is the kbanner utility. Finally you'll need a copy of EDICT, which you can get from the Monash University Japanese language FTP archive. A good mirror of this site to use is "ftp://ftp.cdrom.com/pub/japanese/monash/". To install the library you need to edit the files config.sh, config.h and Makefile. setting the various configuration options there as appropriate for your system. Then type "make install". Hopefully everything compiles and installs correctly. If so test out the utilities by using the newly installed kl utility by doing something like "kl -c setsuna". The libraries and utilities can be configured to use either Unicode or EUC Complete Two-Byte encoding as the wide character type used internally. It is strongly recommend that you choose EUC. While Unicode may appear to be the solution for all the world's communication problems, the only significant effect it has on these utilities is to increase their size by almost an order of maginitude. If you do want to use Unicode you'll first need to get set of three translation tables, JIS0201.TXT, JIS0208.TXT and JIS0212.TXT, which are only available from Unicode, Inc. You can download them using anonymous FTP from the directory /PUB/UNIX/MAPPINGS/EASTASIA/JIS/ on their Internet host unicode.org. The utility pts is a prototype front-end processor (FEP). Unlike other FEPs for Unix it doesn't require (or use) X Windows, instead it requires STREAMS based pseudo-ttys. While only Solaris 2.3 is supported, porting other System V Release 4 based operating systems may be fairly simple. It might be improved one day to support BSD-style pseudo-ttys, however Linux pseudo-ttys don't support the required functionality. In summary, comment out the "MAYBE_PTS=pts" line in the Makefile unless you're using Solaris 2.3 and don't mind using a completely undocumented FEP. If you have any questions, comments or bug reports, please send e-mail me at rridge@csclub.uwaterloo.ca. Ross Ridge