Message-Id: <6.2.1.2.2.20050513141010.06116e40@mail.jefsey.com>
Date: Fri, 13 May 2005 15:06:11 +0200
To: "Antoine Leca" <Antoine10646@leca-marti.org>,
	<unicode@unicode.org>
From: "JFC (Jefsey) Morfin" <jefsey@jefsey.com>
Subject: Re: support of numbers
In-Reply-To: <007101c557a1$9d6859e0$46c86464@arcesa.com>
References: <6.2.1.2.2.20050511210419.0493eb10@pop.online.fr>
 <007101c557a1$9d6859e0$46c86464@arcesa.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Sender: unicode-bounce@unicode.org
Errors-To: unicode-bounce@unicode.org
Precedence: bulk

I am not a linguist. I am a network architecture and brainware person. The 
point you rise is very important. It is the point of the man/machine 
language. For example I search a term right now to charectarise machine 
languages (not software languages): English or French structured to be 
fully understandable and culturally acceptable while being used by 
computers (for example two web services). In reference to 2001 Space 
Odissey I though about HAL as in HA-languages and lool for HA good and funy 
meanings. I suppose such languages will also have their own scripts and 
charsets, etc.

Now, let consider base 8 or 12 or 16 numbering systems. Should we use them 
or not? IPv6 is Base 16 and Telephone Base 12. Computers are Base 2.  As 
you say there are many systems which should all at the end of the day 
permit to support any figure. Should we use them?

What we know is that there around 5 billions of interoperable telephone 
sets (land lines and various forms of mobiles) and related numbering plans 
in use. This means that the system works for: everyone "without distinction 
of any kind, such as race, colour, sex, language, religion, political or 
other opinion, national or social origin, property, birth or other status" 
and therefore can be a common standard for all peoples and all nations, and 
be embedded in technical propositions which will strive to respect the 
human rights and freedoms at national and international level, towards 
their free, secure, stable, universal and effective recognition, observance 
and usage among all people, in the respect of the sovereign law and 
jurisdiction of each States and the natural intergovernance in subsidiarity 
of the granularity of the global, international, national, local, 
structural, personal digital ecosystem and of the continuities of 
intelligent services they may support.

So what selected is not a locale numbering system, but a common used 
telephone pad usage (you alos use when you want a chocolate bar in a 
vending machine). This obviously calls for a deeper study (my question over 
numbers was a first check) like:
- what are the various behaviours of the people using the various scripts?
- dead scripts must be provided with a numbering pad corresponding to their 
current keyboard (or the keyboard adapated)
- addltional characters should be standardised (telephone provides 10-12 
but not in old scripts)
- this must be discussed and user-tested.
- etc. and this etc. is probably where there is most of the effort. This is 
why I documented it and call for remarks.

Thank you for your attention.
jfc









At 11:53 13/05/2005, Antoine Leca wrote:
>On Wednesday, May 11, 2005 9:11 PM JFC (Jefsey) Morfin wrote:
>
> > I need to support telephone numbers (0-9) in an multilingual
> > application.
> > I would like to know if I miss some numbers, if in the Ethiopic case
> > I can assume that the first 10 values are 0-9 and how do Kharoshthi
> > support decimal entries?
>
>Your subsequent explanation was a lot of fun for me. Thanks.
>
>However, there is something I fail to understand. A number of scripts have
>various, overlapping, number systems: in general, one is semi-positional (so
>have the "digits 1,2...9,10,100,1000 etc. No zero) and then it evoluted into
>another, completely decimal (only ten digits). The main examples are Chinese
>of course, and Tamil; I believe the Babylonian and Mayan systems can be
>considered in the same category, but they are yet encoded in Unicode, and
>furthermore they use a 10/6/10/6... resp. 20/19/20/20... bases you'll not
>like.
>
>Then, if you consider using Ethiopian (or Kharoshti) numerations, I do not
>see why you would discard the similar Chinese or Tamil ones... and then you
>will have the problem to decide, o r to let decide, for example for Tamil,
>if one should use the "traditional" system (using 10, 100, 1000) or the
>"decimal" one, using the added zero.
>
>
>Antoine