Re: [idn] Thoughts on nameprep

[Bruce Thomson <bthomson@fm-net.ne.jp>]

I can help out with the Japanese keyboard explanation. When typing
in Japanese you have a choice of several keyboard input packages
available, of which IME is the most popular. Without IME or equivalent,
you would be limited what is on the keyboard itself, which is basically
your standard ASCII. With IME, you are able to type a whole range of
hiragana, katakana, Kanji, and special symbols, all by typing alphabetics
and then converting.

The default input mode is hiragana, which lets you type sequences of
hiragana
by typing the equivalent alphabetics. For example, type r and a to get the
single hiragana character pronounced ra. You can then either confirm the
hiragana character, add additional hiragana to the sequence, or press
the space bar to convert it into something else. It converts it into what
it guesses is my most likely choice, and I can press the space bar again
to get the next choice. Pressing space a third time gives me a menu with
all possible choices.  With "ra", I get 13 menu choices: the original
full-width hiragana, 4 Kanji, full-width katakana, half-width katakana,
and the alphabetic sequences RA, Ra, and ra, in both full-width and
half-width alphabetics. Whichever choice I make, IME remembers it as my
preferred choice, bringing it to the top of the list of choices next
time I convert the same characters.

IME can be quite annoying in its attempts to be intelligent, remembering
previous choices which were exceptions, or typed by someone else borrowing
my computer, etc. So it's certainly possible for it to encourage mistakes
by leading you to input something you didn't intend.

At this point user names and passwords are probably without exception
limited to ASCII, but depending on what software you use to enter them
it might be possible to accidentally type Kanji, kana, or full-width
alphanumerics. So people learn to be careful.

As to why people still might use double-width numbers (or alphabetics,
for that matter), some people seem to feel that they look better
when intermixed with Kanji text (which is full width). Personally
I don't use them, but some people do. The only reasonable use I have
seen for half-width kana recently is on cell phones, because of the
limited screen size. These non-standard characters are not used
for non-text input such as user names, passwords, credit card
numbers, telephone numbers, zip codes, etc.

If someone gave me the opportunity to work on the Japanese keyboard spec,
I would like to consider trying to wean people off these old characters
by refusing to convert to them (this could be a checkbox option). You
can always type them if you desperately need them by switching keyboard
modes. Currently in addition to the default mode I described above, there
are modes for typing katakana, ASCII, full-width alphanumerics, and
half-width kana directly without the conversion step. I expect that
creators of proportional fonts would then work to make mixtures of
alphabetics and Kanji blend more smoothly.

Bruce

"D. J. Bernstein" wrote:
> 
> Yoshiro YONEYA writes:
> > 'intelligent IME'
> eel
> Could you please explain this in more detail? How does the typical
> Japanese keyboard interface work?
> 
> What exactly do you type if you want to connect by FTP to ftp.aol.com
> and log in as anonymous with password yone@nic.ad.jp? What would you
> expect to type if aol and the username and the password were replaced by
> Japanese names?
> 
> Suppose we settle on fast nameprep: it's the keyboard interface's job to
> help you type good domain names, so that other programs don't have to
> worry about bad domain names. What changes would you make in the
> keyboard interface to support this?
> 
> > Half-width kana will be obsolete
> 
> So, if we take the slow nameprep approach, then in twenty years we'll
> have a bunch of networking programs with the useless skill of converting
> half-width kana to full-width kana. Right?
> 
> As for numbers, Bruce seemed to say that most applications expect ASCII
> digits, and that double-width numbers won't work. If that's correct, why
> is anybody using double-width numbers?
> 
> ---Dan