Re: Which type of mushaf ins Unicode encoding?

[A Lynch <al-arabeyes at alinsyria dot fsnet dot co dot uk>]

Thomas Milo wrote:

> Now there was the authoritative source that claims there was originally a
> single vowel followed by a small or big noon. I consulted another
> authoritative source, Dr Gerd-R�diger Puin, researcher into the history of
> Qur'anic orthography, and he confirmed my observation that the oldest
> manuscripts express tanween with two, horizontally aligned vowel signs. This
> is also how Yasin Dutton describes them - no trace of a small noon, let
> alone a big one. Yet, as a logical device, I like the elegance of the
> formula:

> [vowel <a/u/i>]+[any tanween <regular/iqlaab/idgaam>
> (as many as you can identify; these are the three expressed in the Saudi
> orthography, AFAIK)

I agree with that except that the iqlaab also applies to nuun saakin, and idghaam also
applies to nuun saakin and various letter combinations such as laam + raa' (qul rabbii).
Therefore they should not be considered as variations of tanween but as modifiers
of tanween, nuun saakin and other letters.

The other modifier is ikhfaa', which applies to tanween and nuun saakin. Ikhfaa' applies
when nuun saakin precedes a letter which does not cause idghaam, iqlaab or idhhaar:

Final Nuun Saakin or Tanween followed by y/r/m/l/w/n  يرماون  => idghaam إدغام (assimilation)
Any Nuun Saakin or Tanween followed by khaa'/`ayn/ghayn/hamza/Haa'/haa'    خ ع غ ء ح ه => idhhaar إظهار (make clear)
Any Nuun Saakin or Tanween followed by baa' ب => iqlaab  إقلاب (change nuun to meem)
Any Nuun Saakin or Tanween followed by any other letter (seen, sheen, faa' etc) => ikhfaa' إخفاء ('hide' the nuun)

In the saudi mushaf the ikhfaa causes the tanween to be staggered and the nuun saakin
loses its ra's alkhaa', take a look at:

a) nuun followed by raa' (idghaam) => staggered tanween or nuun has no ra's alkhaa'
b) nuun followed by khaa' (idhhaar) => regular tanween or nuun has ra's alkhaa'
c) nuun followed by sheen (ikhfaa') => staggered tanween or nuun has no ra's alkhaa'

So ikhfaa' looks similar to idghaam except that for idghaam with r/m/l/n there is a shadda
on the following initial letter. In the maghribii & pakistani mushaf the idghaam has a shadda
over all the idghaam letters y/r/m/l/w/n i.e. including waaw and yaa'

The letter meem has similar rules and also the script changes in similiar ways.
For instance with idghaam (meem saakin + initial meem) the meem loses it's ra's alkhaa'
and the second meem has a shadda. With ikhfaa' of the meem (before a baa') the meem
also loses its ra's alkhaa. With idhhaar (meem saakin followed by any letter other than
baa' or meem) the meem keeps the ra's alkhaa.

We end up with such as:

<vowel><tanween><iqlaab> <space> <baa'>

and for example:

Anbi'hum bi asmaa'ihim => 
	<hamza qaT`><fatha><nuun><sukuun><iqlaab><baa'><kasra><hamza qaT`><sukuun><haa'>
	<Damma><meem><sukuun><ikhfaa'><space><baa'><kasra><space><hamza qaT`> etc

The first <iqlaab> tells the renderer to display a small meem over the nuun and no sukuun.
The <ikhfaa'> after the meem saakin tells the renderer not to display the sukuun over the meem.

> This structure guarantees searching in existing Unicode-enabled
> environments. It also guarantees that modern font technology can take care
> of the shapes, whether Pakistani , Egyptian, Saudi or North African. This
> approach would mean that on the level of plain text code, Qur'ans remain
> identical when they do not conceptually differ and it would make research
> into real differences much more efficient.

If you treat all the tajweed marks as modifiers and leave tanween as a single variant then
all the tajweed marks can be treated as whitespace and all vowels are consistent whether
in tanween or not. Perhaps it would also be nice to mark sukuuns wherever there is no vowel  and
the renderer omits rendering it when required (over a nuun with idghaam or ikhfaa for instance).

> A simple canonical equivalence insures legacy compatibility with existing
> fathatan, dhammatan and kasratan.

This is your field but it seems clear.

Although I am sure this is how the quran should be encoded at a top level, I am not 
totally convinced that it applies at the unicode level. We need to be sure that any
'automation' of applying glyphs does not stand in the way of any unorthodox use that
the user may require. I guess that this can be achieved by the user not using tajweed codes
but the actual glyph itself. For instance, what if I just want to print a small meem by itself? Should
I be allowed to do that and if so, how? Also, as Gregg says, this could all be hard to explain to
a font encoder.

wassalaam
abdulhaq