Re: [idn] IDNs in email message bodies

[John C Klensin <klensin@jck.com>]

--On Monday, 26 March, 2001 08:30 +0100 Maynard Kang
<maynard@i-email.net> wrote:

> I assume that when you say "translate to Arabic or Chinese"
> you mean it in a linguistic sense (as opposed to doing
> encoding conversion).

Yes.

>... 
> For all purposes, an ACE domain name present in a message
> should be treated as arbitrary data by any translation engine
> (whether human or machine), and hence MUST NOT translate that
> ACE string into Arabic and Chinese together with the rest of
> the message content.

I agree with your reasoning.  But, if domain names are presented
to the user in ACE form as often as I think this implies, we
have invented an ADN (ACE-domain-name) system, not an IDN
(International...) one.

That is probably not good news although I note that similar
problems would occur with most, or all, other systems (including
sending UTF-8 and non-DNS-based ones) that can regularly present
internal forms to users without unambiguously identifying them
with escape sequences or the equivalent.

> But you are right in that there is no way to discern an
> ACE-string from other data with perfect accuracy. Perhaps we
> need an escaping mechanism for this.

Probably.
    john