I'm just saying UTF-8 everywhere is even more unrealistic than any
other options at hand.
Too bad, because it's the only option that's remotely practical in the
long term. Do you really think every programmer who wants to mung text
is going to include code that supports not only the hundreds of extant
character encodings but also the seventeen kinds of in-band and
out-of-band declarations of them?
Even if there was a easily-usable well-supported library to do all
this, and you managed to get everyone to use it, the minor differences
between encodings hidden by its abstractions would lead to subtle bugs
that would continue to plague international users and force them to be
second-class citizens forever.
If you want international users to be on the same level as those who
just use plain ASCII, a drop-in solution is the only way to go, and
UTF-8 is the obvious drop-in solution.
--
Aaron Swartz: http://www.aaronsw.com/
(For the purposes of this email, "international users" are users who
need characters other than those in plain ASCII.)