From: JFC (Jefsey) Morfin (jefsey@jefsey.com)
Date: Thu May 12 2005 - 22:56:41 CDT
On 07:43 12/05/2005, Dean Harding said:
>I wrote a little app which converts between Ethiopic and arabic-indic (aka
>0-9) numbers just for the fun of it a while ago. You can check it out
>here:
><http://www.codeka.com/blogs/index.php/dean/2005/02/04/parsing_ethiopian_numbers>http://www.codeka.com/blogs/index.php/dean/2005/02/04/parsing_ethiopian_numbers
>
>Do Ethiopian telephones really use ethiopic numbers, though?
I thank everyone for help! May be can I document a little bit more the need
for you to understand as it may give ideas. And feed backs.
The application is to be put into a browser, plug-in or any internet web
service access panel or used in PADs (private alias directories: your
internet "phonebook" where you enter the name you want for the IP address
you want in the language you want and which works everywhere for 36 years
this year).
So you may enter the URL 1.707.222.45.67 and get somewhere (preferably on
the site of the owner of the telephone number). But you may also access a
documentation center in using its Interface Grid (the directory of its
canonical[or not] numbers for an information). This is where even
Kharoshthi numbers can be used (you work on a Kharoshthi document using
Hans' smart keyboard: you want to call a reference, you will use Kharoshthi
numbers). The Interface Grids are by nature Hexa (IPv6) but they will
probably be standardized as decimal and >9 reserved for special plans or
service characters, like on the phone (#,*, @, etc.) which works with 12
figures. Except for ISO 10646 Hexa like tables.
This is why the languages "telephonic" scripts are of importance (there are
a lot of charsets). In every language including Kharoshthi. The conversion
Ethiopic to decimal too. Because the canonical IPv6 (Uninum)
Interfacenumber will be the same (pivotal multilingual information) for
everyone, but will not be _numbered_ the same by the users.
Obviously a PAD can also be used and numbers being aliased with tags.
Let consider that Hans has defined a touch-screen keyboard and he can
easily switch charsets (we are in dynamic Internet, so better speak
Internet): he wants to look like a Russian speaking person, using Cyrillic,
of military background chating in a relax manner on the web and having a
good scientific background. He will use the corresponding script
(Cyrillic), language (Ukrainian), regional (Russian), referent
(Military language) and style (casual) subtags and will call an ISP
(intelligent service provider) CRC (context reference center) able to
filter them positively. When he has done that he will select there a
scientific orientation for his context: for jargon, formulas, quotes,
ontologies, calendar, jokes, statistics, etc. He will save his langtag with
the date (in case subtags change definition, like for "CS" in ISO 3166) as
"HansRu" at a given InterfaceID, which will keep the Interface ID of all
these elements for him for some additional services. He can also save
HansRu-one as the context he got on this CRC, but he can go on others for
more, and possibly add them in HansRu-two, etc.
With that anywhere he is, he can call HansRu-one and have all the
information and Word personal dictionary and style parameters entered to
chat with all the assistance specific to the avatar he wants to play.
Obviously one sees that the locale notion attached to the computer and
partly to a region, etc .. will blur and be inserted in the context which
is attached to the user. For a good understanding: the locale is a software
part to control your hardware, the context is a brainware part to control
your software. You can play with locales and you can play with contexts....
and get conflicts :-)
I work on the basic technicallities of Contexts for three years. We created
a national test CRC structure for that. And we start having making big
"suggestions". InterfaceGrids are probably a very important standardisation
issue including plug and play functions, e-home, e-city management, etc.
Another is the dialog with CRCs which is to be quick and fast: direct IPv6
address is good, but XML :-( ... Tagging is interesting as in fact a very
simple, pervasive, proven and efficient support system is .... the DNS. You
enter your personal tag (this permits you to rename the whole Unicode names
in every languages every day) and you get its canonical InterfaceID on
every CRC (an Interface ID is like a telephone extension, but there are
trillons of them, for each use you may have).
Obviously web services, or applications can use them. Let assume you
develop in a Chinese environment and never see an ASCII character. You want
to enter a French langtag for a person of the Versailles area using the
Universalis Encyclopedia as a language reference writing an administrative
document. You will enter each of these subtags aa Chinese labels attached
to the chinese.chinese CNNIC domain name of your CRC's access server (a
simple smart name evaluation system over the DNS - it saves management and
permit entries in disorder) were it may know the granularity of Versailles
from its postcode, or default to France. It will find for you the
corresponding IP, and the numeric encryption for it. The French receiving
end will check the numeric langhage encryption and get all the elements
from its own Latin local CRC tables.
This works for every "human compatible" exchanges. Let consider two
webservices dialoguing in e-Spanish (Spanish version for
inter-human/computer usage, a new language to add to ISO 639-6). The
specification of the language will permit them to dialog, but the user will
be able to follow. Now if you interface an OPES (a process which acts on
the traffic flow - cf. IETF WG-OPES), it will look at the words and will
call on the CRC when there is word of e-Spanish its context does not know:
it will add a footing note to explain it.
jfc
This archive was generated by hypermail 2.1.5 : Thu May 12 2005 - 23:01:18 CDT