Re: Arabic wasla in Unicode

[Miikka-Markus Alhonen <Miikka-Markus dot Alhonen at tigatieto dot com>]

Hi!

On 07-Jan-02 Mohammed Elzubeir wrote:
> Salam Mikka ;)

Wa-`alayk salaam, kadhaalik :)

My name, BTW, is spelled with two i's as well as two k's, i.e. "Miikka", not
"Mikka" or "Miika". I know it's hard but what can I do? :)

>> The specific problem I have in mind at the moment, is fonts. Does anyone of
>> you
>> know of a font which would include wasla on its own, i.e. not on top of
>> alif?
> 
> I think I have an _idea_.. but I may be wrong. This, after all, is a
> transparent glyph,
> right?

Well, it's a transparent character in shaping behaviour, but still it's a
visible one... Normally non-spacing combining marks are placed
above/below/left/right to a dotted circle U+25CC in Unicode charts, so that
they could be recognised easily.

> There are Unicode 'experts' that drop by every now and then -- I wonder where
> they are.
> I would highly appreciate it if you CC' any contacts you make regarding this
> matter.

Sure I can do that, if you like.

> Actually, maybe we should consider the uses of such an addition. I know that
> through-out
> my academic experience, I have never had to have that displayed on its own.
> It is also
> something that will need to have its own mapping to the keyboard, or am I
> wrong?

I guess you have never tried compiling an Arabic primer, have you? :) I know
this character is quite rarely used in real world, but there are some occasions
where it could turn out to be useful.

Actually, what I was thinking, is how Arabic speaking people usually consider
alif mawsula? Is it an undivisable letter like qaf, where the two dots just
are an essential part of the letter (i.e. there is no such character dotless
qaf in modern usage), or is it clearly a letter alif _with_ a separate
diacritic, the wasla? An analogous example in Latin alphabet would be something
like lowercase letter "i", which just can't exist without the dot (except in
Turkish :). OTOH, a letter like "b with dot above" (U+1E03), which is used e.g.
in Irish, is clearly a letter "b" which is somehow modified (a grammatical
feature of "lenition" is applied) and therefore it's not a separate entity.

So, even for purely theoretical reasons, is there a need to encode wasla
separately? As for keyboard mappings, it's probably going to be used so
rarely that an extended Arabic keymap with wasla won't probably need to be
created... When it's needed, something like \u0656 can be typed. Or one can
have some kind of a transliteration map where e.g. "a" = alif, "W" = wasla and
therefore "aW" = alif mawsula (in this kind of maps ligature "lam-alif" would
also be just "la", not a separate key).

Best regards!

----------------------------------
E-Mail: Miikka-Markus Alhonen <Miikka-Markus dot Alhonen at tigatieto dot com>
Date: 10-Jan-02
Time: 0:03:24

This message was sent by XFMail
----------------------------------