Re: mule Chinese input inconveniences

[Miles Bader]

richter@poincare.math.northwestern.edu (Bill Richter) writes:
> Miles, I think that's a lot more natural in the japanese input method
> than in GB, since the input strings are so much longer.  In GB, you
> only type a few characters.  But in the japanese input method, don't
> you type whole sentences?

Not typically (at least I don't); it can sometimes figure out where to
split things if you _do_ convert a multi-word input, but it's not very
reliable, and I find it easier just to do a single word at time.

> Sounds more complicated than GB.  Actually I don't see anything about
> RET to end translation here...

Trust me.

I think that even in the simpler chinese case, it would be good to have
RET exit the conversion state -- it's very important to have a reliable
and consistent `exit this special state now' key (even if there are
other more proper ways to exit).  Since RET is already used by isearch,
&c., and generally is associated with the concept of `entering' or
`accepting' input, I think it would be the natural choice.

-Miles
-- 
Love is a snowmobile racing across the tundra.  Suddenly it flips over,
pinning you underneath.  At night the ice weasels come.  --Nietzsche