On Tue, 24 Feb 2004, Derek Martin wrote:
> ...But I can't help but think
> that it would be better if the kernel allowed for language-specific
> IME modules in the console/tty drivers...
(Assuming I've understood the meaning of "IME" correctly...)
That only solves the easy part of the problem. The hard part remains:
the kernel still has to accommodate large, complex input modules which may
want to consult the filesystem, interact with the user in complicated
ways, etc. This is code which doesn't belong in the kernel *at all*.
(In fact, I would claim that a good bit of the current tty driver doesn't
really belong in the kernel.)
If it's awkward to deal with this entirely in user-land -- which it might
be -- then probably the preferred solution is some kernel mechanism to
interpose a (language-specific) user-level daemon in the kernel input-
handling chain. It's complicated because of the possibilities for user
interaction, though.
> Then you could deal with it
> uniformly at all levels of input management... One API to enter
> characters, whether you're typing in a terminal emulator or at the
> console.
Just to be difficult... :-) Exactly how the user interface for this
should work might actually depend on where you're typing, so uniformity
may not be a virtue. An X environment gives a sophisticated interface
rather more options for user interaction than a text-mode console does.
(How useful those options are... is hard for me to say, since I'm not on
close terms with any of the languages where these issues get serious.
But the potential is there.)
Henry Spencer
[EMAIL PROTECTED]
--
Linux-UTF8: i18n of Linux on all levels
Archive: http://mail.nl.linux.org/linux-utf8/
