Comments on draft-ietf-idnabis-defs-10

John C Klensin klensin at jck.com
Mon Aug 31 12:01:47 CEST 2009



--On Monday, August 31, 2009 11:52 +0200 Elisabeth Blanconil
<eblanconil at gmail.com> wrote:

>...
> Elliptic or not:
> 
> - This WG is chartered to deal with "xn--". Dealing with
> "00--" to "zz--", is a change indeed.

No, the WG is chartered to deal with IDNA.  We are not proposing
to use those prefixes, only to impose a condition on their use
for IDNA-aware applications.  That condition (or some very small
variation on it) has been in the documents since almost the time
the WG started.

> - The same as imposing lowercases to punycode is a change in
> the punycode algorithm and/or ACE (impact on UTF-8
> conversion?). 

No one has proposed changing Punycode, although it has been
tempting.  As I noted, both the ToASCII operation of IDNA2003
and the operations of IDNA2008 produce lower-case from the
Punycode algorithm.   Under IDNA2003, if someone later mixes the
cases, the internal Stringprep processing causes the anomaly now
under discussion to be eliminated.  The question is how to
prevent the anomaly under some edge cases in IDNA2008 and
preserve the existing behavior.

- The same as declaring invalid some domain
> names because of the way they are written is a change in the
> ACE and DNS

Unless you are referring to a case I can't think of right now,
this is not being done.  

> - The same as mapping some characters (unicodepoints?)
> requires the charter change you requested.
 
> Wil is right: the whole thing does not stand well. We all know
> that. We can be incremental, disruptive, messy or smart about
> it. We should stay on our incremental tack. And be smart
> further on (interplus). Disruptive changes on this last
> working minutes is messy. There are plenty of pending issues
> which have not been addressed yet (except by Harald) at
> http://wikidna.org/index.php?title=WG-IDNA_LC. I wish we are
> done with IDNA2008 and get it published in 2010, before I am a
> grand-mother.

Regardless of your intentions, your behavior pattern is not
advancing things.

    john




More information about the Idna-update mailing list