https://bugs.documentfoundation.org/show_bug.cgi?id=137716

Hossein <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[email protected]

--- Comment #12 from Hossein <[email protected]> ---
(In reply to MohammadReza Hosseini from comment #4)
> (In reply to Ming Hua from comment #3)
> > I know nothing about Persian or Arabic, but it seems from a Persian MS
> > Office user [1] that Persian and Arabic indeed have slightly different
> > alphabetical order, and MS Office exhibits a similar problem for Persian
> > users.
> > 
> > 1.
> > https://answers.microsoft.com/en-us/msoffice/forum/all/persian-alphabet-or-
> > arabic-alphabet/1930e610-6439-4858-94bc-833aad8c61b5
> 
> I checked the Arabic alphabet and you are correct. In Arabic alphabet the
> letter 'و' comes after 'ه'. So this bug only affects Persian language.
The point is exactly this difference, and LibreOffice locale-specific behavior.

Arabic and Persian share some characters, but they order is different. Then, in
Persian locale, sort is done according to Persian order of characters, in
Arabic and other locales including English, sort is done according to Arabic
order of characters.

Is this a good approach?

Let's see the parameters:

1. Locale settings: application locale (and system locale)
   + Set in "Tools > Options > Languages and Locales > General"
2. Default languages for Documents
   + Set in "Tools > Options > Languages and Locales > General"
3. Language set for a text
   + Set in "Tools > Language" and visible in status bar

Language UI (4) is different from above. The locale string shows 1, 2 and 4.

Locale: en-US (en_US.UTF-8); UI: en-US
         ^^       ^^              ^^
        app      system           ui

UI (RTL/RTL) and text / paragraph direction (LTR/RTL/Top to bottom) can be
important, but not here.

The question then becomes, what should be the behavior to select the correct
sorting behavior.

Currently, it is based on locale. That could be good if using Persian locale OS
and application (LibreOffice) was common. Unfortunately, it is not common at
all, because the translation for OS and also LibreOffice itself is incomplete.
Therefore, people use Persian (and probably also Arabic) language with English
locale, as you see here.

Also, even if everyone was using their own language for the OS and LibreOffice,
by using locale-specific behaviors, multi-lingual text processing becomes
limited.

You may want to have a column in Persian, and another in Arabic. This can be
separated by the language of the text in Calc, and also in Writer and
elsewhere.

Therefore, I argue that text language is better for sorting. This approach has
its own limitations, for example if you mix multiple languages in a column.

It makes sense to test the same thing in other office applications. Below is a
discussion about a similar issue in Excel and Power Query:

Sort Persian alphabet is wrong in Excel and Power Query
https://learn.microsoft.com/en-us/answers/questions/5139119/sort-persian-alphabet-is-wrong-in-excel-and-power

One additional note is that this difference between Arabic and Persian alphabet
order also exists for آ and ا. It is visible in the pictures from above
discussion.

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to