If ACE wanted to be liberal with what it accepts, it would not
insist that applications "MUST" stop with an error when it finds
that the encoded string has an ASCII representation. Political
decisions about uniqueness should not require everyone to have
to upgrade their servers to software that thinks it knows which
of the Unicode table entries correspond directly to ASCII glyphs.
For what applications would ACE be superior to UTF-5?
Is there an Internet Draft describing the differences between
the two?
Cheers,
James