Re: [idn] I-D ACTION:draft-ietf-idn-idna-08.txt

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

--On Tuesday, 11 June, 2002 00:19 -0700 Dave Crocker
<dhc@dcrocker.net> wrote:

> At 12:48 AM 6/11/2002 -0500, Eric A. Hall wrote:
>> Now that I have answered your questions twice, perhaps you
>> could answer mine. What is the exact problem you foresee with
>> allowing an application to put whatever it needs into the RRs
>> it uses?
> 
> The simplest answer is that IDNA has nothing to do with
> putting anything into RRs.
> 
> IDNA is about a micro layer above DNS and below applications,
> as it says quite explicitly in the IDNA document.
> 
> Any other concerns are secondary.
> 
> You seem to be insisting on pursuing a very interesting topic
> but in the wrong forum.

Dave,

I started a much longer note yesterday in response to some of
your questions, then concluded that the detail, and perhaps the
whole note, was a waste of people's time.  But, since the
questions seem important, let me try to summarize the questions
and then try to address the issues.  (I believe Eric has
identified all of these issues and that I am just presenting
them in a different way, but that may not be correct -- I may
have misunderstood some of his remarks in a way that has
inspired other thoughts.)

To dispose of the last question first, yes, I find that close
variants of this discussion have been going on for two weeks, and
we still aren't sure what the question is, to be very troubling.
Given the possibility that there might be an important topic for
the committee to deal with, I sat out most of those two weeks,
listening and trying to understand the issues.  I think that is
a responsible approach, although you could (and I think have)
argue that, at this late date, anything that shows up is
sufficiently likely to be irrelevant that shutting down the
discussion is more efficient.

One result of the confusion is, I think, that we have managed to
get four separate topics intertwined and conflated.  That is
never a good way to make progress, so I want to try to separate
them back out.  In semi-random order, they appear to be:

(1) Is IDNA properly and adequately specified for the uses to
which it will be put and is it designed appropriately for that
range of uses?  It seems to me that this question is relevant to
any specification that IETF is considering standardizing.  The
answer may depend somewhat on the answer to the next question.

(2) Is IDNA closely bound to internationalization of the DNS or,
as you note above, is it really "about a micro layer above DNS
and below applications" and hence "has nothing to do with
putting anything into RRs"?  And are Eric's concerns properly
addressed to the IDNA specification or should they be focused
somewhere else?

(3) Should the applicability of IDNA in the DNS context, and the
xxx-prep procedures to be used with it, be specified on a per-RR
(and per-Class) basis?

(4) To what extent should out-of-band communications between
applications, which utilize strings which the applications might
construe as internationalized domain names, influence the design
of IDNA and, if so, what should the impact be?  I have problems
with this one, because I think it leads us straight into a
rathole, especially when character sets and representations used
in applications (or on particular systems) get into the mix.
But, especially if IDNA is a micro layer for applications use,
rather than a DNS adjunct, it seems to me that questions about
the implementability, usability, and efficiency of a proposed
protocol have always been fair game for discussion wherever the
protocol is discussed.

I'm addressing the first three of these in separate notes.  I'll
leave the fourth to others, but ask that they separate it into
another thread so that people who don't care can more easily
ignore it.

    john