Keith Moore writes:
Yes, it's probably desirable to migrate to a new format which relies more
on utf-8 and less on ascii-compatible encoding schemes.
The American Standard Code for Information Interchange is more than
simply the set 0, 1, 2, ..., 127. It specifies an _interpretation_ of
each of those numbers as a (printable or control) character, and thus an
interpretation of strings of those numbers.
RFC 2047 is _not_ compatible with ASCII. RFC 2047 specifies a different
interpretation of some of those strings.
In contrast, UTF-8 is compatible with ASCII. Every ASCII string has the
same interpretation under UTF-8.
The feature of the RFC 2047 encoding that you care about is that it's a
7-bit encoding. In particular, it's compatible with some buggy MTAs. In
contrast, UTF-8 requires some MTAs to be fixed.
You and the other IDNA GoofyCode people make yourselves sound like fools
when you misuse the phrase ``ASCII-compatible'' to mean ``7-bit,'' and
in particular when you refer to blatantly incompatible encodings as
``ASCII-compatible.''
---D. J. Bernstein, Associate Professor, Department of Mathematics,
Statistics, and Computer Science, University of Illinois at Chicago