[Idna-arabicscript] mapping of Full Stops

[John C Klensin]

--On Monday, October 12, 2009 09:57 -0700 Erik van der Poel
<erikv at google.com> wrote:

> On Mon, Oct 12, 2009 at 9:49 AM, "Martin J. Dürst"
> <duerst at it.aoyama.ac.jp> wrote:
>> On 2009/10/12 1:03, Erik van der Poel wrote:
>>> U+3002 has a much longer
>>> history in IDNA and is much more firmly established. If
>>> U+06D4 would be mapped to U+002E at the data interchange
>>> level (think HTML), there would be a period where IDNA2003
>>> implementations and new implementations would resolve domain
>>> names differently.
>> 
>> Was U+06D4 allowed inside a label in IDNA 2003? If yes, then
>> this will indeed be a problem. If not, then the difference is
>> only between being resolved (with appropriate mapping in IDNA
>> 2008) and not resolved (in IDNA 2003), which is an
>> unfortunate but tolerable difference in implementation
>> behavior.
> 
> U+06D4 is allowed inside a label in IDNA2003. Sorry, I should
> have been more explicit about that.

Sorry, I hadn't even thought to ask that question.  My apologies.

If it is valid intra-label in IDNA2003, I don't see how to make
a recommendation about this --in Mapping or elsewhere-- without
revisiting all of our discussions about differences in
interpretation of strings under IDNA2003 and IDNA2008.  I hope I
don't need to remind anyone that each such affected character
has led to extended and heated discussion, claims of creating
security problems, and requirements to document both the
security issues and transition strategies.

Maybe this situation is different from our friends final Sigma,
Eszett, and the differences in interpretation of
formerly-ignored characters in those respects, but I can't see
how.

   john