From ALB@immedia.ca Sun Jun 17 14:20:00 1994
Received: from Clouso.CRIM.CA by dkuug.dk with SMTP id AA05462
  (5.65c8/IDA-1.4.4j for <i18n@dkuug.dk>); Thu, 16 Jun 1994 21:15:10 +0200
Received: from immedia.ca by clouso.crim.ca (4.1/SMI-4.1)
	id AA12950; Thu, 16 Jun 94 15:14:58 EDT
Return-Path: <ALB@immedia.ca>
Received: by immedia.ca (3.2/2.D)
        id AA4409; 16 Jun 94 19:21:05 +1900
Date: 16 Jun 94 19:20:00 +1900
From: ALB@immedia.ca
Message-Id: <199406161921.AA4409@immedia.ca>
To: bealle@torolab6.vnet.ibm.com, cpwg-mail@revcan.ca, paref@vm1.ulaval.ca,
        umavs@torolab6.vnet.ibm.com
Cc: i18n@dkuug.dk, sc22wg20@dkuug.dk, tc304@dkuug.dk
Subject: Decimal delimiter and thousand separators: slight distinction
X-Charset: ASCII
X-Char-Esc: 29

----------
Baldur (from Iceland) says, in answer to my posting:

>It would certainly be logical to have a special decimal delimiter, and also
>perhaps a thousands delimiter, for the Latin/Greek/Cyrillic (and perhaps
>other) scripts. This would be congruent with the decimal and thousands
>delimiters already implemented in 10646 for Arabic.

It is certainly required for presentation and is indeed part of POSIX
LOCALEs.  However thousand (or any general chunk of digits) delimiters are not
absolutely required on input and for massive data entry, it is generally not
allowed as it diminishes productivity (well, this is debatable, but it is not
absolutely essential).  But delimiting the fraction from the integer part of a
number is a universal function.  The point is that it has no reason to be
dependent on the presentation parameters which are part of the LOCALE.  It is
the way it is done now, though.  That is, in my humble opinion, very incorrect
for any user interface, and, as I demonstrated, can be counter-productive in
multilingual environments.

I would agree with Baldur, however, that the end-user has less odds to
make mistakes if he/she enters numbers by chunks (of 2, of 3, or whatever).
But that does not require a key to be depressed (it is a matter of personal
rythm of work and method), unlike the decimal delimiter function which does
require it.

Alain LaBont<e'>
