----- Original Message -----
From: "Keith Moore" <moore(_at_)cs(_dot_)utk(_dot_)edu>
To: "Edmon Chung" <edmon(_at_)neteka(_dot_)com>
Cc: <ietf(_at_)ietf(_dot_)org>; <idn(_at_)ops(_dot_)ietf(_dot_)org>
Sent: Thursday, March 21, 2002 5:32 PM
Subject: [idn] Re: Moving Towards UTF8 vs ASCII(ACE) Forever
second, "ASCII for the rest of our lives" is a mischaracterization.
IDNA allows applications to accept and present IDNs in native
form, without requiring all applications and infrastructure to
upgrade before IDNs can be used. IDNA is designed to maximize
the rate at which IDNs can be deployed.
I totally disagree with this, how can the deployment of IDNA speed up the
rate of IDN deployment, given that IDNA has to upgrade all client
applications or OS with a plug-in and deploying server-side solution for
example at one ISP, or TLD will serve users in the 1000s.
users don't care whether IDN queries are encoded on the wire.
they only care about whether IDNs work for them. IDNA puts the
user in a position to determine whether IDNs work for him -
a UTF-8 everywhere approache forces him to wait for everybody
else to adopt IDNs before he can use them.
That's why without a hybrid approach, IDN will never work. Especially IDNA,
it has leakage(user cares because the ACE name is not the name they want to
pay for!!), requires a plug-in, or future change in the application(which
means not working nowadays even if we decide to deploy this).
My own thinking is that the R&D effort of adding the ACE conversion to
future applications is the same as making protocol change in future
application. However, if you look at the transitional stage, using IDNA MUST
require a plug-in, using a hybrid system will make the transition smoother,
where most of the existing system works without plug-in and if future
applications sends out new protocol will work as well.
If you look at this as a simple math equation of what you get now and what
you get in the future:
Methods Now Future
ACE+IDNA ~(x/100M)% 100%
Hybrid ~95%* 100%
(where x is the number of net user willing to install client plug-in, and
roughly around 100M net users nowadays)
*(from our official test results)
At least you can get an average of (100+95)/2% throughout the deployment
for Hybrid Systems, whereas using ACE in order to achieve the same result,
can anyone of you tell me what x needs to be: 95M!! I thought we are
engineers and scientists and this should be obvious enough and scientific
enough to prove why we need a hybrid system to make the deployment of IDN
more smooth(and Think for the users!!).
David Leung
Chief Technology Officer
Neteka Inc.
T: (416) 971-4302
http://w!.neteka.com