With just a few small changes I was able to build a Windows 2000 version from the same source code. It works almost exactly the same as the Windows CE version. The menu bar is different, the application can be resized. Since there is a keyboard the enter key does a look up in EDICT and some control keys do other searches.

This program should display Japanese characters, provided the MSGothic Japanese font has been installed. In order that other application display Japanese characters the Windows 2000 region settings may need to be set with Japanese as the default setting to read and write documents. For Japanese input the Japanese IME need to be installed. The fonts and the IME all comes on the original Windows 2000 CD so its easy to install.

Download the exe file and the cut down kanjidic from here and the latest EDICT and ENAMEDIC with the index files from The Monash University Nihongo ftp Archive at "http://ftp.monash.edu.au/pub/nihongo/".

When the program first runs it expects the dictionaries to be in "C:\Program Files\" so use the dictionaries dialog box to change each of these to the real location of all the dictionaries. This is a bit messy and something I have to work on.

This version does not work on Windows 95, 98 or ME as these version don't support Unicode. I have not tried it on Windows NT4.0, it might work with the Japanese font installed, but probably it will just display junk.

Dictionaries

The program now accepts dictionaries in EUC code or Unicode, UTF-8 or UTF-16 little-endian. For Unicode files the first few bytes of the file must be the standard Unicode BOM (Byte Order Mark), anything without these bytes is assumed to be in EUC code. UTF-8 BOM is EF BB BF, UTF-16 little-endian BOM is FF FE. For information on Unicode see "http://www.unicode.org".

EUC code uses 1 byte for English characters and 2 bytes for Japanese, this gives the smallest file size. UTF-8 is a little larger if the files has only a few Japanese characters. It uses 1 byte for English characters and 3 bytes for Japanese characters. UTF-16 uses 2 bytes per character, so it is significantly larger than EUC but can be smaller than UTF-8 if there are a lot of Japanese characters in the file.

The best source of dictionaries, information and utilities is probably The Monash University Nihongo ftp Archive at "http://ftp.monash.edu.au/pub/nihongo/".

The format of the dictionaries and optional index files is the same as for EDICT. The applications like ejoin (for joining dictionaries) and jdxgen95 (for making the index file) can also be used to create new or customized dictionaries in EUC code. There are also programs to convert to and from other codes such as UTF-8, JIS and Shift-JIS.

Mike Johnson