[At-Large] IUTF FAQ preparation

JFC Morfin jefsey at jefsey.com
Thu Jan 12 01:47:54 UTC 2012

I will try to collect the questions raised and answer to produce and 
maintain an IUTF FAQ.

At 17:47 11/01/2012, Lutz Donnerhacke wrote:
>You only need to register the CLASS according to RFC6195 Section 3.2.

Correct. However, this calls for an agreement between the different 
parties involved as the IDNS belongs to everyone. Compatibility is of 
the essence, as a use need, not as an engineering need. Therefore use 
(IUTF) must have its say as much as engineering (IETF) and possibly 
Govs if the ICANN/GAC was acknowledged by the IGF.

>Then you need to distribute an additional "root hint" file pointing to your
>new root servers (similar to http://www.internic.net/domain/named.root)
>As you can see, they only claim to be responsible for ./IN.
>Then you can start selling domains from your new root.

The target is not to start selling. It is to document properly the 
way the entire IDNS is to work to the benefit of the whole digital 
ecosystem (WDE) users. Once this documentation is clear, stable and 
accepted there will certainly be profit and non-profit registries and 

However, the IUTF has identified the necessity for IDN (International 
digital names) to be life-long for URI stability for exemple, TM 
protection, personnal identity right, etc. IDNs are not specially 
related to the Internet. Therefore the registry job in this context 
is different, and most probably legal. The, DNS and ML-DNS services 
will call for operators to support zones and most probably new services.

Also, the development of IPv6 (as a young protocol) will most 
probably lead to new development. In particular I would support the 
notion of "environment conversion", this would mean to consider that 
by default the Internet+ is non-profit. However, if a commercial 
(declared as such) host is involved, it may change the environment to 
for-profit, with new rules applying. This is a notion to certainly 
look at within the neutrality context.

>Please allow me to add a professional advice. If you try to convince the
>browser manufacturers to ask for your class instead of IN, you might have a
>real market.

The IUTF role is purely architectural. It is to document the way the 
Internet+ works should work for users to more adequately benefit from 
it. The difficulty with current application manufacturers is that 
they believe that when the user uses their application, the 
application is the center of the world. And they want to be smarter 
than the user, guiding him/her with their ideas. Actually, what the 
Internet+ means is that the center of the world is the user: 
he/she/it decides of the context in which the application operates. 
Building that network context is not the job of the appcation but of the IUI.

The browsers are to be transparent as every other application. They 
have no responsibility in determining the language, the CLASS, the 
script, the encryption, etc. This belongs to the Presentation layer 
and is the job of the single preDNS function in the ML-DNS [ML-DNS is 
actually a smart front-end of the Internet DNS].

So the issue is the architecture of the ML-DNS as the *single* smart 
interface between UTF-8 transparent CLASSless applications and the 
ASCII/CLASS context. With possibly its own multiclass root system.

The matter is a matter of power and size. The architecture must be 
able to scale, i.e. to support millions of TLDs. This is not going to 
be done with the RSS and 96% of erroneous requests. The requests must 
be locally addressed, so the root is to be local and only list the 
IRN (International Root Names) the user is interested in, hinting the 
Root Name Servers by the operator he favors.


More information about the At-Large mailing list