[Idna-arabicscript] mapping of Full Stops

[John C Klensin]

(duplicative copies trimmed)

--On Tuesday, October 13, 2009 23:04 +0500 Sarmad Hussain
<sarmad.hussain at nu.edu.pk> wrote:

> However, do note, in context of other comments, U+06D4 is a
> DISALLOWED code in IDNA200x protocol.  This does imply that
> any previous registered domain names with U+06D4 as part of
> the label would be rejected by any application conforming with
> the upcoming IDNA protocol

Yes.  And this may be one of the situations in which an
incompatibility between IDNA2003 and IDNA2008 is an advantage,
at least in the medium to long term.  U+06D4 as a hyphen
look-alike is problematic (as is hyphen as a U+06D4 look-alike,
as your example pointed out).  Having domain name labels with
U+06D4 embedded is just a problem --phishing or otherwise--
waiting to happen.

On the other hand, if we reach the point a few years down the
line that most software is following IDNA2008 rules so that
U+06D4 is getting purged out of zone files and rejected where it
does appear, it should then be easily possible to say "U+06D4
can sensibly be locally mapped to a label-separator if that is
sensible for any character at all".  It is just hard to make
that recommendation on a global basis right now.

> Thank you all for the fruitful discussion. 

Thanks for your patience in working with us in the discussion
and arguments.

best,
   john