Carl,
> I have lamented the lack of a good IME interface to capture ruby as the text
> is entered. If nothing else they can be useful for some types of sorting.
But in case of e.g. Japanese, this would definitely not work out in every
text. If a particular word is not included in the dictionary of the
respective IME, the user has to enter the characters another way , for
instance by entering compounds containing the desired character and deleting
the rest. But this way, the reading entered does not necessarily have to be
identical to what was entered!
Example: I want to enter "kokken" (land + to sharpen/polish), being the
abbreviation of "kokuritsu kokugo kenkyuujo", the "National Institute for
the Japanese language" (s.th. like that). Well, it's not in the dictionary
of my IME. Therefore, I enter "kuni" (another reading for "land"), then
"kenkyuu" ("research") and delete the "kyuu"-part, there you are.
(especially names are entered using this method)
If you recorded the ruby automatically, this would lead to nonsensical
results. (*kuniken instead of "kokken").
Of course, this is primarily a Japanese-specific problem (due to its
multiple readings of characters). But even in Mandarin Chinese, there are
AFAIR a few characters with two readings: e.g. the character for "big",
being "da" in most cases - but "dai" in "daifu" (physician).
Regards,
Berthold
