Re: [RAM] Re: Renumbering impossibility: TSL/SSL certs, DNS delegation etc.

[JFC Morfin <jefsey at jefsey.com>]

At 20:08 08/08/2007, Gert Doering wrote:
> Guess why IPv6 (with "dedicated networks of incredible size!!") wasn't
> a huge success with end users?  Because they *don't care* about how
> the network works, and whether they have a static for dynamic IP address.

I am really sorry Gert, but this is not the case. This is a recurrent 
disagreement I have with the "technical" side of the TF, and a 
support from the user side (and leadership). The reason why the users 
are not interested in IPv6 is because they have no incentive to use 
their addressing part because it is not protected by any standard, 
what shows them the IETF has not understood their need so it is not 
an exciting thing to consider as it is not supported. And ROAP is the 
proof. You are concerned by the number of address for the network, 
they are concerned by the number of their own addresses and how they 
can distributed them across the world.

I know we disagree. Even Thomas who asks for inputs for his summary 
does not want to take into account an UP (User Protected) addressing 
area. This is why I ask if we want more than a patch. (But I also way 
that the current technology under a different paradigm should support 
the need, because the need is for more and more addresses, not for 
traffic). This is why I do not mind much about the solution you will 
find - all I need is virtual bandwidth we can organise with the true 
addressing we need. Obviously this will inelegantly and costly pile 
unnecessary layers ...

I work on semantic addressing, distributed translation memories, 
referential systems. Address grids are certainly every user most 
demanding IP¨v6 address applications I know - right now. Billions of 
metadata having its own ID in thousands of languages (have you only 
considered what the RFC 4646 or a semantic network may require in 
terms of hardware access to be workable).

This is where the IETF decentralized paradigm is totally outdated. At 
most the entire network may need 30 billions of user addresses. This 
is very small when compared with plug and play multilingual semantic 
sub-addressing grids on a family SNHN (small network home network). 
SNHN are the core of the Internet. And you doubt they only need a 
single IP ..... Please think distributed fractal networking, and stop 
thinking decentralized host access supporting unified Google 
information in a single language and a single presentation layer. 
That was academic test 30 years ago.

You certainly heard about what is an ontology and a thesaurus. The 
largest taxonomy is probably a Dutch dictionary with 450.000 words. 
Add all the names of all the people, all the possible conceptual 
adjencies, etc. This is what every user need just to talk with his 
own computer, like Hall in Space Odyssey, at a low cost and 
acceptable robust rusticity. Not in 3010, but in 2001 (we are 
probably 15 years late). Translation memory/semantic addressing call 
for each Semantic Being to have an IP address. An ontology is built 
by a person using a taxonomy to describe her vision of the world. I 
suppose that a few years from now you will have your own and your own 
knowledge base on your mobile.  We still are on the same planet, but 
in different worlds.

Cheers !
jfc



_______________________________________________
RAM mailing list
RAM at iab.org
https://www1.ietf.org/mailman/listinfo/ram