Yeah, I'm surprised and impressed by this article. When I was living in Japan last year I looked into the "Chuditch" Iyonix project but never took up the offer in the end.
As to modernising RISC OS, full Japanese input support would be several times harder to implement than supporting Unicode, the character set and the necessary fonts, as the article suggests.
On Windows the Japanese IME is very fully featured - I enter text in romaji on a standard UK keyboard which it then converts into hiragana as I type, and then substitutes the correct kanji after I press Return. To tell the IME that I am entering katakana (e.g. for my name) I'd have to go into the IME and switch to Katakana. Japanese users would do things differently with hiragana symbols on each key on their keyboards, saving the first stage of the conversion process, and pressing the hankaku/zenkaku/kanji key (usually located just under Escape) to toggle between half-width/full-width and kanji characters, and using Shift to enter katakana. Sometimes there are twenty or more conversion options for a given hiragana entry and to be effective the IME must present the most commonly-used options first - it is a pretty complex beast! At the end of a year of typing Japanese almost daily and still didn't understand all the options available in the Windows IME.
It looks like we are well on the way to having Unicode, the character set (UTF-8, although most Japanese emails I receive are iso-2022-jp) and hopefully the necessary fonts, on RISC OS 5 at least. That should result in the ability to read emails in Japanese and view Japanese webpages on RISC OS at some stage in the future.
I wonder how well the Drobe Comments system copes with Japanese text