However, none of these allow entry of Japanese text in any convenient form, nor any real support of Japanese in general.

So I am developing modules and applications which will allow RISC OS to have Japanese and Chinese input modes / fonts etc.

Does this sort of thing interest ANYONE else???

Post back here preferably, or email me on willblatchley@yahoo.com.

Should it matter if there's an demand for your desired feature? I say write it for yourself, while being safe in the knowledge that people will love you for making RISC OS even greater than it is now.

'allow RISC OS to have Japanese... fonts' If I'm reading that correctly, you're saying you can get the font manager to handle (converted) ttfs like cyberbit. I thought the whole reason that the people behind Oregano use their own font manager was that the RISC OS one couldn't handle 'complex' fonts. How are you (planning on) coping with this problem?

Then again I could be completely wrong.

I thought that the problem with the RISC OS font manager was not that it couldn't, theoretically, handle the fonts, but rather than it doesn't support unicode formats and as such it's limited to the number of characters that can be represented.

Then again I could be completely wrong.

I guess he will be really pleased to hear that...

Why not ask them if they are ahead or behind you?

I am working more on the input side of things. I am not so much concerned with the font output (I also heard a long time ago that there was a Unicode compliant font manager around, waiting to be released). However, since Japanese and Chinese have many thousands of characters, you can't just "type" it in on a keyboard like you do with a Roman script. Instead you type a Romanized equivalent of the Japanese/Chinese which my software will aim to convert into the correct script.

There are many stages to this:

- Converting the Romanized text into Japanese Kana (Phonetic script).

- Converting some of the Kana into Kanji (Ideographic script, almost identical to Chinese)

- Converting that foreign script between various character sets and encodings (Shift-JIS as used by Windows, UTF-8, UTF-16 and UTF-32, the Unicode encodings etc.)

- Plotting it!

It is all made complicated by the fact that there was no (until Unicode came along) common standard bringing together all the different encodings. If RISC OS was to fully implement Unicode, I think it would be a major plus point for it.

Looking forward to hearing more opinions.

Will

As for direct Kanji input... I have not seen that form of IME on any platform before. And I think there is good reason. In Japanese, there are many words which have completely different meanings, but exactly the same pronounciation. So in any case, there needs to be some way for the user to interactively chose which word he/she is intending to type. This is usually implemented by a using a pop-up window, and cycling through the different kanji options using a key on the keyboard (often SPACE). Once the right one is found, it is selected (usually with ENTER).

Even in Chinese, the user would normally type in pinyin (romanized Chinese), and then have that converted into the correct Kanji characters. There again, there are often several different kanji corresponding to the same pinyin.

Nevertheless, what you say could be done, I guess. The user would type a whole string of Romaji or Pinyin, and then it would be converted (at some point in time requested by the user, or automatically) directly into Kanji and Kana. But I think it would be confusing for most people to see a long string of Romaji. It is generally much easier to read Kana I think.

There are other input methods to consider, like mouse-controlled character drawing pads (these often seem to employ neural-network based, or wavelet-decomposition based recognition algorithms). The one on Windows AtoK13 (by JUSTSYSTEM) is, I have to admit, very impressive! But these are too slow to substitute keyboard entry in my opinion.

Try rspooner@argogroup.com - although he's recently got married and might not be back at work yet.

There are many stages to this:

- Converting the Romanized text into Japanese Kana (Phonetic script).

- Converting some of the Kana into Kanji (Ideographic script, almost identical to Chinese)

Wouldn't it be better to convert direct from keyboard characters to Kanji, and then figure which to convert into Kana? The reason being, this would be more efficient if adapted for Chinese. Better still, make it configurable via a menu setting.

btw, I recall Pace were advertising some vacancies a while ago