Re: [ga] Re: Root server traffic

[JFC Morfin <jefsey@xxxxxxxxxxxxxxxx>]

Dear Andy,
the reality of naming is of a TLD forest. At least a three roots 
forest: the ntia root and .local + .chaos.
The ICANN ITD TLD testing is a pure technical joke. No one ever 
doubted that a root server system could not support a few TLD more 
:-)  and wikipedia shows everyday that a wiki can support more than 
17 languages.
Everyone serious agreed a long ago (and documented last year in 
Athens [Vint and Falström] that IDNA RFCs had to be revamped before 
they could deploy seriously (please indicate where is the WG-IDN? May 
be you will remember that Brian Carpenter [IETF Chair until last 
year] was a former Unicode BoD Member, and that his predecssor was 
Harald Alvestrand, Unicode and ICANN BoD Member.
Chinese names show that the IDNA's problem is not with the basics of 
the IDNs, but with phishing (3rd level and higher), with the 
permitted character set and the choice of Unicode, etc. and that has 
nothing to do with "IDNccTLD" but with Unicode being supported, with 
the TM problems related to "babel names", with the political control 
ICANN and the USA want to keep on the internet and the linguistic 
support (which is the key of the semantic processing and of 
e-commerce [e- for Engish inside]), with Affilias or Verisign or etc. 
wanting to sell names in China or in India etc.
There is so much a long time that IDNA works that I do not even 
recall in the http://jean-françois.jefsey.com page :-).
On 20:24 23/11/2007, Andy Gardner said:
> This proves that CNNIC HAS added extra TLD's running parallel
> alongside the "approved" ICANN root.
>
> Why has ICANN never said anything about this?
>
> I understand that many ISP's outside China have added them as well,
> to cater for the Chinese people in their community? Tiscali?
A few years ago I provided free (no registration) Internet access all 
over France (through telephone calls) to every known TLD.
> Verisign's "Global Digital Brand Management Services" actually
> announced they were selling these TLD's in their news bulletin...
>
> http://gnso.icann.org/mailing-lists/archives/council/msg02929.html
>
> which was later edited to remove the evidence that they were TLD's.
> Quickly swept under the carpet.
The Verisign cooperation agreement with the USG prevents them to 
cooperate with open roots. This made them to lose a lot of money!
> What with the Arabic split root as well, it's clear that IDN TLD's 
> have been tested for quite some time already, so why the need to re- 
> test them again?
They test their control over the ccNSO and GAC, and their 
desinformation capacity which is the real basis for their domainer "industry".
An "IDN" is a regular domain name starting with "xn--": in most of 
the ccTLDs you can register them. You need some smoke screen to make 
the people believe it has something paticular (except that "xn" in 
Europe means "Danger", and in many countries in means "Christians" :-) )
> Add to that, ICANN's iTLD test breaking the "no variants allowed"
> rule requested by the CJK community (which Verisign follows) one must
> wonder just what the hell is going on here.
They want to find what they can permit, because there is no way for 
gTLD to support script/languages. Just remember that IDNA is about 
supporting the whole Unicode character set on shot, in writing them 
in ASCII (punycode). Not languages, nor scripts. Each ccTLD may claim 
that they only support the characters (2ND level and TLD) used in 
their own script(s) [without any limitation on 3rd level]. But gTLD cannot.
I proposed to consider linguistic/script vitual zones (having 
linguistic TLDs for example - but that is killed by the way Unicode 
and Google wanted to control languages in RFC 4646). They fired me 
for opposing their mess and keeping them somewhat interoperable for 
the time being. But they could not fire the problem :-)
> Can any country run a spilt root now?
Well, the problem is that it was easy to create Chinese names as a 
single extension. It is more difficult to run several ones without 
agreeing on a technical procedure. So, before we can really have a 
domain name forest system (DNFS) we need someone to publish best 
practices, so ISPs or individual users (with a local resolver) know 
how the other people proceed and make the same, or propose a better solution.
Now the question is: would that be better? would that be a mess? etc. 
The only thing I know is that it would create competition between 
ICANN and others. If this is for money there is no real interest, if 
it is for user based open use vs. ICANN walled garden?
jfc