Re: IAB letter on IDNs and VeriSign
John C Klensin <klensin <at> jck.com>
2003-01-28 08:37:13 GMT
--On Tuesday, 28 January, 2003 13:27 +0900 Soobok Lee
<lsb <at> postel.co.kr> wrote:
> <quote>
> [14] At the core of all of the IAB's concerns is the
> architectural principle that the DNS is a lookup service which
> must behave in an interoperable, predictable way at all levels
> of the DNS hierarchy. Furthermore, as a lookup service it is
> such a fundamental part of the Internet's infrastructure that
> converting it to an application-based search service, as the
> deployed system does, is not appropriate even in the case
> where the query presented would not normally map to a
> registered domain. </quote>
>
> I have to accept IAB's concerns about VGRS solutions. But, if
> such solutions will be timedout and phased out after 2 years
> and operated with care under IAB's new "operational
> guidelines", those harms can be minimized, while it will give
> great boostup to plugins distributions. I expect ICANN will
> consider the bright side of the solutions and make balanced
> decisions. Many asian region NICs are considering to put
> even non-ASCII labels in their zone (not IDNA-compatibile)
> with slightly different approaches, even after they become
> fully aware of those caveats behind their approaches.
> IDNA/stringprep/nameprep introduce confusions,collisions and
> uncertainty to the DNS systems much more than VGRS specific
> implementation/deployment solutions does ,as explained in
> ICANN BoD's last resolutions in Shanghai..
Soobok,
With the understanding that I'm not on the IAB any more and did
not participate in the writing of that statement...
First, designing anything for the Internet on the assumption
that it will be deployed and then "phased out after 2 years"
just doesn't recognize reality. We have had almost zero success
in "phasing things out" when they are deployed only on servers.
The problem is much worse when implementations require replacing
software on client machines and the protocols are still vaguely
operational. Our last success in making a major and rapid
phase-out solution was the "phase out" of NCP. That was 20 years
ago and with the benefit of credible threats to remove any host
from the network which did not comply.
Second, there is no reason to predict that ICANN will do
anything that is sensible, measured, and that shows a good
understanding of either the technical or the policy issues of
IDN deployment -- rather than just the short-term politics or
agreement that internationalization is good. Certainly the
resolutions adopted in Shanghai should not be considered good
news in this respect. Those resolutions basically say "there is
a problem, we should proceed carefully, and the lifetime of the
so-far-ineffectual committee should be extended to work on it
without changes that will make it more effective".
> IESG/IAB are
> willing to justify that IDNA by supplemental Registrations
> guidelines ? If IDNA can be justified in such a way, why not
> VGRS solutions ?
Still I hope IAB can express such
> consistent conservatism also to IDNA/nameprep itself, which is
> the source of all these hassles...
At the risk of reopening a debate that was never satisfactorily
resolved, if the issue is merely one of permitting non-ASCII
characters to be used in identifiers, with those identifiers
being placed in DNS labels, IDNA is fine. It may even be more
complex and protective than is needed for that purpose (just as
the LDH rules were more than is needed for strict identifier use
in ASCII). To my knowledge, neither the IAB nor the IESG have
taken a formal position on whether registration guidelines are
either necessary or sufficient to resolve any presumed problems
with IDNA. If, by contrast, someone is expecting IDNA to solve
the most general of user-interface searching and navigation
problems, that expectation is completely unrealistic. I think
just about everyone knows that is the case at this stage, if
only because it is clear that the DNS is inadequate to solve
that problem for names that are restricted to ASCII. Some
marketing organizations might like to pretend otherwise, but
that doesn't make success in use of the DNS as a general-purpose
search environment any more likely.
So your point was?
regards,
john