OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index]

On Wed, 16 Dec 2020 at 13:50, Roger L Costello <costello@mitre.org> wrote:

Hi Folks,

Unicode make it possible to write things in different languages.

For example, rather than this XML:


a Bengali-speaking person can write this:


Or, in a programming language, rather than this assignment statement:

              Number_Students = 42

a Bengali-speaking person can write this:

              সংখ্যা_ছাত্র = 42

That’s awesome.

But, but, but, … how come that universality doesn’t extend to digits?

How come we can only use these digits: 0 (hex 30), 1 (hex 31), …, 9 (hex 39)?

Why, for example, can’t a Bengali-speaking person use the Bengali digits: Bengali digit 0 (U+09E6), Bengali digit 1 (U+09E7), …, Bengali digit 9 (U+09EF)?

Why, for example, can’t a Bengali-speaking person create XML such as this:


or write a program assignment statement like this:

              সংখ্যা_ছাত্র =

Let me explain why I assert that the Bengali-speaking person “cannot” do that.

Numbers in an XML document or in a program are just strings and, to perform arithmetic operations on them, those string numbers must be converted to actual numbers. I looked at the source code for the C function (strtol) that converts strings to numbers and here is the key to how it converts a character digit to a number digit:

              digit_number = digit_character - '0’


That generates a number digit by treating the character digit as a number and subtracting the number corresponding to the character ‘0’. For example, if the character digit is ‘4’ (hex 34) then when we subtract ‘0’ (hex 30) we get the number 4. Perfect! But ……… only if we allow European digits (0, 1, …, 9). Clearly, if we were to subtract ‘0’ (hex 30) from the Bengali digit 4 we do not get the number 4.

Thus I conclude:

  • When expressing numbers, the only digits that can be used are the European digits
  • Unicode is universal, but that universality does not apply to digits or numbers

I don't see that the conclusion follows from the earlier comments. Clearly you can convert various number string forms to  numbers, just because strtol doesn't do that doesn't allow a universal conclusion that it is not possible.

It is a choice in the language design what syntax to allow for digits, just as in keywords.  in xslt for example, you have to use for-each  as the element name, you can't use a Bengali (or French) name; it just is what it is.

You say that that you could write this  <সংখ্যা_ছাত্র>42</সংখ্যা_ছাত্র> but if the input is following some standard schema for student records the chances are that isn't valid and it has to be <Number_Students>42</Number_Students>   You could write a filter to take the localised form, mapping the element names, but any such filter could mp the digits at the same time

You could allow multiple language keyword interfaces and number syntaxes, but in most programing languages (for good or bad) the choice is not to do that at the lower level programming end of things and only offer localised interfaces at a higher level
So a spreadsheet might accept localised digits, and localised date forms "Wednesday 16th December 2020" but at some lower level the date might need to be in  standard  form 2020-12-16 or whatever.


Obviously I am not understanding something correctly. Please help me to understand.


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index]

News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 1993-2007 XML.org. This site is hosted by OASIS