Re: [idn] What's wrong with skwan-utf8?

[Brian W. Spolarich]

On Sun, 24 Dec 2000, D. J. Bernstein wrote:

| idn-compare says that this ``doesn't work with many exiting [sic]
| protocols that relies [sic] on DNS.'' Which protocols? What exactly is
| so difficult about allowing UTF-8 in those protocols too?
| 
| A few people on the mailing list have commented that particular pieces
| of software would need to be changed to handle UTF-8. So change them!
| The total number of lines of code is small. Nobody has come up with a
| proposal that would give users IDNs with less overall effort.

  UTF-8 is an 8-bit, variable length encoding of UCS-2.

  The various ACEs are (essentially) 7-bit encodings of UCS-2.

  Using 8-bit data will break some applications.
 
  Using 7-bit data (presumably) will not.

  One objection that many people have to ACE-based solutions is the
'leakage' problem, where users wind up seeing ACE-encoded data in contexts
where they expect to see the name as they typed it.  However, a
significant number of applications and OS localization environments do not
know how to deal with UTF-8 either, so this tradeoff is probably moot.

  -bws