Folks,
Regarding the MU encoding, there are a few issues that still need to
be addressed (of course!).
I intend the following as a starting point for discussion, rather than
my "final" answer. These are some of the issues:
Although network protocol standards often do not need to deal with
issues of screen display, in this case we probably need to. There is
an unwritten(?) rule that says that users' terminals are 80 columns
wide, and people generally write their messages in such a way that
each line fits in 80 columns. Most people would agree that each ASCII
and each Latin-1 character takes up one such column, and most users of
the Japanese language would also agree that Japanese Kanji take up two
columns each. When the MU encoding is used together with a Content
Type of "text/plain", we probably want to be able to assume that the
receiver will see on the screen what the sender intended him/her to
see. It would be a lot of work to specify a column width for each
Unicode character, so perhaps we should just specify widths for the
"obvious" characters (ASCII, Latin-1, Japanese and maybe others), and
then just say that the rest of the characters have a maximum width of
two columns.
Although the encoding allows ANY 16-bit code to be encoded, the
specification must prohibit Joe Schmoe from using any old 16-bit code
on the network. I.e. we must refer to a particular version of the
Unicode Standard.
If there are any "unstable" parts in the Unicode Standard, we need to
identify them and possibly recommend that people don't use them until
those parts are "fixed". For example, I have heard that the Thai
national body is unsatisfied. Is there a possibility of an
incompatible change in the future?
That's it for now...
Erik