Comments on the Unicode Codepoints and IDNA Internet-Draft

[John C Klensin]

I agree with the general comments that Ken, Patrik, and others
have made, but let me add two things:

(1) Our criteria for what we permit and prohibit have changed in
the last year.   This is something that doesn't affect us going
forward but may be helpful to Jaeyoun and his colleagues in
terms of understanding the difference between discussions a year
ago and discussions today.  A year ago, we still had the "maybe"
categories.  A classification into "MAYBE NO" would have been a
warning to a (non-KR) registry that they should think carefully
and obtain advice before registering names containing these
characters (if I recall, the archaic scripts were in MAYBE NO as
well).  With those categories gone (for several good reasons),
we really need to look at harm, as was suggested in this thread
and yesterday afternoon.

(2) In today's environment, the right solution to the "clueless
other registry" problem is probably to do what was done for
Chinese in RFC 4713 and that is now in progress for Arabic --
publish a document that offers advice for appropriate and
inappropriate characters for registration and, as needed,
explanations of those choices.   That way, if a registry is
interested in advice, they can find it (and maybe find pointers
to more).   Those who are not interested won't, but then we are
back the apparent absence of harm... and trying to regulate
against stupidity (where "stupidity" in this case is a comment
about my personal opinion about registries (at any level of the
tree) who permit domain names in scripts about which they have
no understanding and seek no advice that would improve their
understanding).

     john