Re: Mapping Hindu Numbers for ISO8859_6 and Windows Cp1256

From: Malek Boualem (
Date: Thu Feb 18 1999 - 22:49:26 EST

Omar Halaseh wrote:

> There might be an error with mapping number 0-9 digits (0x30-0x39) for
> ISO8859-6 into Unicode.

No it was not an error but a choice.

> These numbers are in the ASCII sections below 0x7F and should stay as they
> are for any code that manipulates digits in string to work.

They stay in the same position in Unicode (U+0030 to U+0039).

> Currently the ISO8859-6 mapping table maps the 0-9 digits (0x30-0x39) into
> the Hindu numbers in Unicode (0x0660 - 0x0669).

We may not consider that characters in range U+0030 to U+0039
and characters in range U+0660 to U+0669 are the same.
You can use either Arabic or Hindu digits since your text or your
information uses Unicode. The unification principle of Unicode do
not allow to have different codes for the same abstract entities (digits
in this case). Moreover the Unicode standard did not define "escape"
sequences for switching between languages, that was fundamental
with ISO-8859-* series.

> However if you compare with Windows Arabic character set Cp1256 the mapping
> table correctly mapps 0x30-0x39 into 0x0030-0x0039.

I think this is a (wrong) local arrangment since these codes (U+0030 to U+0039)

do not really correspond to Hindu digits.

> Can you please advice?
> Isn't it more proper for the Hinu numbers to be moved from the 0x0660-0x0669
> section into the Arabic Presentation section 0xFE7-0XFEFC since they are
> just a different shape from the ASCII 0x30-0x39 digits?

That does not change anything with the problem you are araising.
It will just consider Hindu digits as secondary characters which is not true.

Malek Boualem
New Mexico State University, Computing Research Laboratory
Box 30001, Dept 3CRL, Las Cruces, NM 88003, USA
Tel Office: (1) 505-646-5711
Fax: (1) 505-646-6218
E-mail :

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:44 EDT