Message-Id: <6.1.2.0.2.20050317074148.0326b670@mail.jefsey.com>
Date: Thu, 17 Mar 2005 08:19:38 +0100
To: Erik van der Poel <erik@vanderpoel.org>
From: "JFC (Jefsey) Morfin" <jefsey@jefsey.com>
Subject: Re: [idn] stability
Cc: idn@ops.ietf.org
In-Reply-To: <42390CCF.1030409@vanderpoel.org>
References: <421B8484.3070802@vanderpoel.org>
 <421D8411.9030006@vanderpoel.org>
 <p06210208be4390618c81@[192.168.0.101]>
 <421E0D0C.2000309@vanderpoel.org>
 <p06210202be43c3888991@[192.168.0.101]>
 <E07CE813AD23B2D95DA0C740@scan.jck.com>
 <421E30F2.1040408@vanderpoel.org>
 <0E7F74C71945B923C52211F3@scan.jck.com>
 <421EA0C9.1010500@vanderpoel.org>
 <00a401c51af3$7863aae0$030aa8c0@DEWELL>
 <A574CA1BE87BFDA3C2A1AC0E@scan.jck.com>
 <42322CE2.4040509@vanderpoel.org>
 <4232B2FD.1080104@vanderpoel.org>
 <4232BA56.5090001@vanderpoel.org>
 <iluk6odazwb.fsf@latte.josefsson.org>
 <00e801c528a8$99ad37d0$72703009@sanjose.ibm.com>
 <ilull8qb5n5.fsf@latte.josefsson.org>
 <42367B63.6080300@vanderpoel.org>
 <4237450A.9010901@v.loewis.de>
 <423754F3.50405@vanderpoel.org>
 <ilumzt47ezc.fsf@latte.josefsson.org>
 <42375C9E.8040001@v.loewis.de>
 <4237917D.9080507@vanderpoel.org>
 <6.1.2.0.2.20050316125429.04464370@mail.jefsey.com>
 <423876D2.2070809@vanderpoel.org>
 <6.1.2.0.2.20050317033553.031631d0@mail.jefsey.com>
 <42390CCF.1030409@vanderpoel.org>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Sender: owner-idn@ops.ietf.org
Precedence: bulk

At 05:51 17/03/2005, Erik van der Poel wrote:
>JFC (Jefsey) Morfin wrote:
>>The issue is easy to understand you need two keyboards to enter the IDNs 
>>and two languages to spell them on the phone - except when the ascii TLD 
>>can be typed.
>
>Some countries have keyboards with both ASCII and local characters. Which 
>countries are you talking about? In any case, we can't just jump from the 
>current DNS to one that does not require typing some ASCII characters in 
>one go. I think these things must evolve, over time.

Dear Erik,
the "two keyboard" formula is a common nickname of the IDNs in non ASCII 
countries.

>>Registries understand nothing to the complex signals (or lack of) from 
>>the ICANN, IETF, market, etc. I am on the WG-IDN of the ccTLDs as a 
>>lingual organization (Eurolinc). I suppose there is one year I received 
>>the last mail. 2 mails in one year++ on ICANN IDN list. They even told me 
>>James Seng does not attend all the ICANN IDN meetings (James?) anymore.
>
>The ICANN guidelines for IDNs probably should be updated. People are 
>talking about this. Hopefully it will happen quite soon.
>
>>James sold us something brillant in MdR2000. But the IETF delivered 
>>something else.
>
>Are you talking about the following?
>
>http://www.dnso.org/dnso/notes/20011112.ICANN-mdr-IDN-JamesSeng.ppt

This looks 2001. I mean 2000. It was a great time when many things came to 
a decision.

>>PADs are obviously the core of the problem and a revolution. Handles 
>>could also be used.
>>In case you forgot about PAD I documented several times, it stands for 
>>Private Alias Directory. This is the Directory of the ML keywords or DN 
>>you can register in your own system to call an external host. This is a 
>>Quick and Dirty excellent solution when an IP address is available. What 
>>will be the case with IPv6 and that main sites can obviously already support.
>
>Maybe people will manually create a few private aliases, but I doubt they 
>would bother to do so for the many, many sites on the Internet. People 
>visit many sites, one after the other, in quick succession. Maybe I'm 
>misunderstanding you...

I am afraid you missed the point. I work on PADs, for security reason. Even 
if there is a DNS dysfunction you keep going through. For the time being 
this is used for security. I documented that at length in Nov. 2003 when 
preparing meetings on nations vulnerability to the Internet. When our tools 
are ready and tested we will have to release them. But I am sure nice kids 
will make them running before. Anyway, this will be like RSS tags on web 
pages. You click and you have the keyword entered in your PAD. This gives 
the site an edge on others: a free keyword in 7260 language, 20000 dialects 
and an infinity of nicknames. Far cheaper than Real Names.

The main problem is to keep them someway consistent with the other name 
space systems.

>>There are several revolutions right now. Just look at the network, the 
>>behaviors and the people. A revolution does not mean an immediate tide. 
>>It means that something we did one way can be done another way, that it 
>>has started being used and that it will only grow. IDNA is not a revolution.
>
>I don't see very many signs of a PAD revolution. Am I in the wrong country 
>(USA)? The wrong mailing list?

You could come on our mailing list. It is in French. But we will sell 
expertise and may be a product to pay for the study. ISP will offer the 
service. I suppose we will publish lists of PAD infos. But this comes to be 
interesting with IPv6. And competition about IP numbers. NRO will love it. 
NRO first provider of ML keywords :-)

>>Naming was left to right first. ARPANET naming was flat and for practical 
>>reason became right to left. It stayed that way. Your proposal works. 
>>Left to right and / separators. I understand there are already 15 
>>millions of them. Named handles.
>
>I don't see them. Do people exchange handles via email? Please show me a 
>handle.

Ask Library of Congress etc. This is Bob Khan's system.
RFC 3650, 3651, 3652. IMHO you are going to love it!

>>A revolution is that less than 50% of the Internet connections still use 
>>the DNS,
>
>Are you saying that less than 50% of the TCP connections on the "Internet" 
>(open Internet or also behind the firewall?) are immediately preceded by a 
>DNS lookup? Where did you get this number from? Please state your source.

Sorry come in the real network land, Europe :-)
This information is probably available everywhere. The authoritative source 
I have is the Chair of AFNIC using notes prepared by Stephane Bortzmeyer. I 
use to trust Stephane.

>>that handles develop, that PADs permits private keywords and ultimately 
>>billions of roots.
>
>Are there any specs for these handles and PADs? Are there any working groups?

Handles is CNRI. IESG expressed a comment which says "not IETF consensus, 
yet it works". Same for PAD, I will not bother writting a Draft, I will 
leave the market using it. Look what they did with our RFC 920 MoU :-)

>>Just a question: when was that the last time you supported a registrant 
>>for his server to be accessed with the DN he just registered?
>>This do help to understand what the people want and do.
>
>Do your registrants use PADs? Are they happy with them?

You kidding? PADs are free.
This is the problem for a Registry!

But this is also something users tend to like. Also that it works. Aslo 
that is simple. Even if Microsoft does not help the programing. But old 
Hosts.txt works well as a starter.
jfc





>Erik
>