Re: [I18nrp] Confusion among characters and strings

[John C Klensin <[e-mail]>]

Nico,

I can't spend any more time on this thread other than finishing
a response to Adam's from Wednesday afternoon.  I've been
working on that intermittently when I've been able to find time
and energy.  It is clearly up to Peter and Pete, who have drawn
the short straw, but I think this notes and the forthcoming one
may be suggestive of some ways the BOF should probably be shaped.

I find your comments about UTR#46 very problematic.  I suspect
they may explain some of the other communications issues we have
been having and the problem is not disagreements about the
specifications.

There are fundamental architectural differences between UTR#46
and IDNA2008 (and PRECIS), differences that became much more
obvious as UTR#46 evolved from what was supposed to be a
transition document from IDNA2003 to IDNA2008 (plus or minus a
few character exceptions that may be important but are not
architectural except as somewhat different philosophies playing
out) to today's document that allows (or urges) several
characters that have never been valid under IDNA2008 or IDAN2003
and that pose possibly-significant issues as identifiers.

Without debating the reasons of who is right and who is wrong
right now (I believe the current version of UTR#46 demonstrates
a fundamental lack of understanding about the DNS and identifier
usage the interacts it, but the authors obviously disagree), the
IETF made its choice among those fundamental architectural
decisions when it adapted IDNA2008 and later carried basic parts
of its architecture into PRECIS.   Nothing prevents you from
believing that UTR#46 is the right model and that the IETF
should get with it.  Nothing prevents you from generating one or
more I-Ds that explain that we made the wrong decision and how
we should transition from where we are now to a model more like
UTR#46 and that specifies that model (IMO, UTR#46 is not
actually complete as an IDN specification, but YMMD).  But your
suggesting ways to adjust, clarify, etc., specifications based
on the RFC4690-IDNA2008-PRECIS model when those suggestions are
rooted in rejection of that model seems to me to be very close
to a DoS attack, even if an unintentional one.

best,
    john


--On Thursday, June 14, 2018 12:00 -0500 Nico Williams
<[e-mail]> wrote:

>...
> As to IDNA... it seems that UTR#46 is non-grata here, though I
> think UTR#46 is correct, and IDNA2008 is wrong where it
> differs from it.
>...