Re: [Bug 54697] New: ALEF MAKSURA shows up as a square at the middle or at the start of a word

[Roozbeh Pournader <roozbeh at sharif dot edu>]

On Wed, 19 Feb 2003, Mohammed Yousif wrote:

> Yes, I saw these two glyphs but they simply have nothing to do with
> Arabic. There are a lot of glyphs in "Arabic Presentation Froms B" as
> defined by unicode that have nothing to do with Arabic. It doesn't
> matter if the glyph I'm seeing is a square or any other symbol that is
> un-related to Arabic (even if unicode says that this is for
> AlefMaksura). An Arabic user will not like the square glyph, but he/she
> of course will be even more dissatisfied if he/she sees a weird glyph
> instead of the square. So it will not help to provide the user with
> un-appropriate glyphs for non-existent letter forms.

0. I don't get you. I have already explained that these have something to
do with Koranic Arabic (I can search and find the Ayah and Surah number
for you). Some Arabic experts say that Koranic Arabic is not at all
related to modern Arabic, and some say that only Koranic Arabic is
important. I don't believe any of them. Both are as important as they can
be. What if some pour soul encodes Koran correctly in Unicode and then
sees that he can't use KDE to display it even when the font and everything
is in place?

1. KDE needs to have some fallback support for displaying glyphs, so
it can show you the glyph if some other font has it. XFree86 has that, and
so has GNOME. This is necessary in multilingual computing, and standards
like CSS2 require it. (Or perhaps KDE has the support but it needs to be
configured properly.) I'm sure the 10x20 glyph is *appropriate* as a 
medial ALEF MAKSURA, and if KDE can fallback on that, ...

3. If a user want an ALEF MAKSURA in the middle of a word, and wants in in 
final form, and doesn't want an space after it, she/he should use ZERO 
WIDTH NON-JOINER.

4. Of course we are not talking about reinventing Arabic standards from
scratch, are we? Believe me that the Unicode model for Persian is also
faulty in some cases and much more problematic than Arabic, because it is
base on the model for another language (namely Arabic). But I evangelize
Unicode compliance heavily. Why? Since it helps us work together, and make
pieces of software work with each other. I hate to see a text file
displayed differently on Windows and Linux, so I fight in both fronts to
make sure both are standard compliant. And then if we find the standard is
faulty, we fix the standard and tell all to copy that.

5. Fonts need to follow Unicode. Ask your font vendor to fix the problem 
for you. If you use Microsoft fonts only, ask Microsoft then. (I know that 
new development versions of MS fonts contain the glyphs.)

6. You are of course allowed to patch your own copy of Qt or redistribute
your patched version. But doing this in the main version, I don't like it.
(Specially since it is our code, Behdad made Qt's Arabic shaping Unicode
complaint and added Persian support. ;)) This is not a bug.

7. Come and discuss this on the Unicode mailing list please. Others may be 
able to convince you and show you the point and be successful where I'm 
failing.

roozbeh