Comments on draft-ietf-idnabis-defs-10

[John C Klensin]

--On Monday, August 31, 2009 10:37 -0400 Andrew Sullivan
<ajs at shinkuro.com> wrote:

> Perhaps what is needed in the mappings document is a
> suggestion that, if the input label is in fact mapped, the
> _mapped_ version of the input label is what is returned after
> lookup?  This would cause the input "Fältström" to be
> returned after lookup as "fältström".  (I don't know if I
> like this idea, but I thought I'd suggest it anyway.)

Andrew,

<editor hat=off>

While several of the other suggestions lead down the path to
this one, it is part of the reason why I'm (slightly) in favor
of a ban on trying to look up or otherwise use A-labels
containing upper case characters.  Independent of
standardization and specification technicalities such as Unicode
independence and a rule-based, rather than table-based, model,
the main innovation of IDNA2008 as seen by the user is a change
to a "what you see and what you get are the same" model from
"what you see and what you get are somewhat similar and
dependent on rules and mappings you probably won't understand".
The latter is the friend of confusion, phishing, and other bad
things.

Without these changes, the definition of U-label is clear, the
definition of A-label is clear, the two are obviously symmetric,
and we don't have issues of
U-label-except-with-some-upper-case-characters.  Encouraging the
latter muddies, and will probably require review, of the other
definitions.

I note that, even under IDNA2003, there were several
conversations in the browser community about showing only what
we now call U-labels --the canonical form-- to the user
regardless of what was actually supplied.

I am quite concerned about Vixie's case-randomization proposal
in this regard and believe that the interests of
internationalization (now and in the future) may be grounds for
opposing it.

But, if we conclude that, because of that proposal or for other
reasons, we should apply lower-case operations, then I think we
need to see those operations as "get a variation into canonical
form and keep it that way" and not "provide an approved
alternate form for a U-label" (or an approved alternate form
that we struggle to preserve").

     john