Japanese Input Method

Japanese needs the most difficult/complex input system.
It has both phonetic (Hiragana/Katakana) and ideographic (Kanji).
There are many ways to input Japanese, though.
Analyzing input data needs translation dictionary
ex) Anthy has 18,000 code line + 3MB dictionary.
FreeWnn has 4MB dictionary.
Windows IME has over 60MB dictionary!
Inputting characters all over the world can use "Finite State Model".
But mechanism is different between Japanese and Chinese.
There are lots of regions to develop!