[discuss] Boundaries and sovereignty

Sun Feb 2 17:11:13 UTC 2014

At 22:16 01/02/2014, Michel Gauthier wrote:
>At 19:47 01/02/2014, JFC Morfin wrote:
>>It consists in providing each person and utilization a full command 
>>on their relational behavior in continuity with their command of 
>>their decision process through their operating system, 
>>applications, and databases. This means, atop of the OSI pile, a 
>>network oriented OS continuation of Posix.
>
>This is certainly an interesting concept, but is that not a lot ambitious?

There is no ambition other than to address the user rather than the 
developer needs through the convergence of what has been planned (but 
not completed yet), exists, is proven (running code), and accepted 
(living mode).

The current interim/working EDP model is based on the master/slave, 
client/server, and cloud models where the network has the purpose to 
support end to end datagrams. Posix is introduced at 
<http://www.opengroup.org/austin/papers/backgrounder.html>http://www.opengroup.org/austin/papers/backgrounder.html. 
This is a computer centered model, not a people centered model as 
demanded by the WSIS.

The real distributed model is P&P where "peer" stands for port, 
process, person, or people depending on the considered stratum. Each 
peer is to be able to organize its own haze (use oriented distributed 
virtual capacities) and relational spaces the way it wants. To do 
that, it needs to master all the functors capable to operate 
morphisms in this haze. Posix defines these functors when the haze is 
restricted to the host.

The internetting project means to support this target (cf. IEN 48), and so far:

- defines the hosts to hosts protocols - ("internet" stratum)
- does not yet support the multitechnology OSI presentation layer 
six, but locates it at the fringe (RFC 1958) or considers their 
applications (e.g. Web, expected ONS for the Internet of Things) - 
(this can be called the "intertech" stratum)
- documents extended services (i.e. active content oriented) oriented 
functors like IDNA (RFC 5895) and OPES (RFC 3835) - (what can be 
called the "interuse" [IUse] stratum)
- does not support semiotic intellition (intelligent information) 
functors - (what I call the "Intersem" stratum that I am mostly 
interested in).

The way I propose to proceed, in order to merge the computer and the 
network model in a unique model, is in identifying and documenting 
their middleware as an Intelligent Use Interface on its own (for the 
network, it is a host virtual front-end or a cloud smart gateway; for 
the host it is the part which puts the haze that it belongs to in 
local continuity).

1. no change at the network

2. no change in Posix, as it's a command set based on the following 
principles 
(<http://www.opengroup.org/austin/papers/backgrounder.html>http://www.opengroup.org/austin/papers/backgrounder.html)

- Application-Oriented
- Interface, Not Implementation
- Source, Not Object, Portability
- The C Language
- No Superuser, No System Administration
- Minimal Interface, Minimally Defined
- Broadly Implementable
- Minimal Changes to Historical Implementations
- Minimal Changes to Existing Application Code

This set already includes a lot of network oriented commands such as: 
endnetent, gethostbyname, sftp, ssh, etc. 
<http://man7.org/linux/man-pages/dir_all_alphabetic.html>http://man7.org/linux/man-pages/dir_all_alphabetic.html

3. an extension and OpenUse publication of this Posix command set 
implementing a full IUI control, and including OPES as an external piping.

This may include commands like dig, nslookup etc and opes, idn, web, 
semantic, wiki, ccn, orented etc. command sets in order to process 
networks distributed applications as a protocol (ASAP) applied on 
"intelligrams", i.e. formated (active) datagrams. Exemple a Wiki 3.0 
intellipage (standalone wikipage).

>For years, considering a Netix Posix continuation could have been at 
>odds with the network stability. For many reasons, including the 
>awareness embodied in this list, that is no longer the case. This is 
>why I will call for a NetixBarCamp this July at the RMLL ( 
><https://2014.rmll.info/?lang=en>https://2014.rmll.info/?lang=en) to 
>consider the status of the art and a roadmap and the HomeRoot 
>project example ("give back everyone their root's data").
>
>
>So, if I understand you correctly you want to explore how to unite 
>local and distributed processing and to start by federating the 
>namespace on a MS concerted basis? But is this possible?
>This would put boundaries and sovereignty under people's authority?

ICANN is an architectonic layer violation. They try their best, but 
this will not change that their functor is about centralizing a 
distributed space and regulating a domain which is and must be MS 
concerted. As long as it was an attempt in a proof of concept this 
was acceptable: it has shown that the "ICANN code does not run" 
(hence the debates). So, it is a technical bug to adequately fix. Up 
to now, we had a brainware limitation (the debate on this list has 
shown once again): mending it without some common awareness was 
risky. OpenStand and Snowden have changed that situation: the 
brainware is aware and busy.

This limitation is the confusion between a file and its data. The 
file we copied to the internet in 1984 was for information. It 
reported the current data that were MS-set-up by the different 
network operators, identified by their DNIC. Postel made it a network 
administration rule (IANA) for his network. Truly globalizing the 
IANA means to give it back its information role about what is really 
happening on the network. The very idea of "balkanization" is 
indecent and it harms the 60 million people living in a place where a 
broad part of human civilization comes from. It only translates the 
inability of the IANA cartographers to follow the life of the world. 
They confuse the map and the territory.

The root is made of the data of the top zone. The root file is the 
anachronic ICANN vision of the ICANN/NTIA "IN" class. It is 
anachronic because hosts in the top zones are declared by their 
fellow TLD name servers and are not registered in the root file, and 
because several TLDs are not recorded by ICANN.

In order to correctly document and keep the top zone clean, we would 
need a permanent survey of the top zone consistency and reality, in 
calling each name server and asking it for the list of the TLDs of 
which they are currently keeping track of in their buffers and the 
associated name servers.

jfc  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://1net-mail.1net.org/pipermail/discuss/attachments/20140202/3a4c03e6/attachment.html>