Issues

1. Character encoding in SHIORI dictionary and setup files
   • Most shiori.dll, especially kawari, assumes that the setup and dictionary files are encoded as Shift_JIS
   • Inconsistency of MBCS charsets
     • Shift_JIS : 0x8140-0x9ffc(except 7f), 0xe040~0xeffc(except 7f) - 2 byte char
     • EUC-KR : 0xa1a1-0xfefe - 2 byte char
     • EUC-JP : 0xa1a1-0xfefe, 0x8ea1~0x8edf - 2 byte, 0x8fa1a1~0x8ffefe - 3 byte char
     • ISO-8859-X : all the char are single byte
   • How can I make Korean (or Chinese, German) Dictionary using current shiori.dll?
2. embryo always asks response for Shift_JIS

   • Charset: Shift_JIS in SHIORI/x.x

   • How about Charset: EUC-JP or Charset: UTF-8

   • Can the embryo convert shiori charset (e.g. utf-8) to local charset?
   • Can the embryo select appropriate font for given charset?
3. SSTP Charset
   • UTF-8 encoding conveys Korean or Chinese successfully? (embryo always changes utf-8 to SJIS)
   • Not only charset, but also Lanuguage is important (Does Japanese want to hear Korean gossip in Korean language?)
4. embryo GUI
   • Is there any catalog or plugin structure for non-Japanese menus or messages?
   • Not touch the main logic. Only resource changes.

Techniques for I18N

1. POSIX locale related functions for charactersets
   • setlocale, ctype.h, wchar_t, mbtowc, wctomb

     Thread UNSAFE!!!

   • OS dependent
2. GNU gettext for message catalog
   • Menu issues
3. C++98 locale class
   • facet
   • Not yet implemented on G++