There are a few Latin typefaces that have extended characters sets, and good diacritic support to cover the needs of scholarly works. For example Brill, and Charis.
What is needed is a good Arabic typeface that is a companion to these Latin typefaces. I list here some properties that would be nice to have. Many of these are in the spirit of the earliest Hijazi Arabic scripts, and which I feel make them happily suitable for the difficult requirements of complete diacritic support. Essentially this would be a "simplified" Arabic typeface.
Images of the Hijazi script below are from the Birmingham Quran manuscript: https://en.wikipedia.org/wiki/Birmingham_Quran_manuscript
Perhaps re-use i's dot in Arabic typeface?
Have similar stroke-widths?
This, I imagine, would be a very difficult requirement to meet. But it would be very nice to have for inline Arabic within Latin text.
Perhaps have bigger counters (aka loop/eye) and smaller descenders. Perhaps get inspiration from Fresco Arabic:
Fedra Arabic also matches Latin vs Arabic leading while going for same readability: https://tptq-arabic.com/fonts/fedra_serif_arabic
Also get inspiration from some design choices of Boutros Text, Mubassat and Maqroo typefaces?
Should ideally be able to render Hafs Quranic text.
Perhaps target all glyphs that Kitab font supports.
Some diacritics can be positioned above or after a basic consonontal character, depending on the writing style and the intended sounding out. Examples are dagger alef and hamza.
Here is dagger alif above a consonant:
And here it is after a consonant:
Hamza similarly can be above or after a consonant. The issue is discussed in detail here:
https://github.com/adamiturabi/arabic-inline-unicode/blob/main/index.pdf
Perhaps allow for both CGJ and Tatweel based positioning in order to render existing encoded Quranic text correctly. Perhaps re-use code from Kitab font.
This again is a tricky one. Diacritcs should be large and spaced out enough to be readable. Yet, ideally they should not require extra leading. And ideally not add noise to basic consonontal letters.
Perhaps:
- Make them grayscale?
- make alif and intial/medial lam shorter when they have a diacritic?
Simple.
Makes it easier to avoid diacritic clashes.
Only one lam-alef ligature: لا
Same lam-alef ligature used for isolated and final form:
Should have enough space between lam and alef for an inline hamza.
Makes it easier to avoid diacritic clashes.
May need care so that waw and raa are not too far visually from next character.
Ties in with only one (lam-alef) ligature. Except:
Text: سيجعل
This doesn't necessarily have any practical benefit.
This is more of a personal preference for the typeface to match earliest Hijazi Arabic scripts.
Text: يستغيثوا
There is a mis-conception that Arabic's final/isolated kaf should be ك and that ک is Persian. The former is only a design choice for naskh and similar scripts.
But ک is simpler and more consistent with initial and medial forms.
Again goes back to earliest scripts and avoid modern tendency of dal and jeem having a very similar shape.
Perhaps get some inspiration from Fresco Arabic's kaf. Drop most of the arm (ذراع) for dal.
Optionally, medial kaf and dal would be identical (like the earliest scripts). But this could be a user choice.
In Naskh the inital bump of final yeh is the previous character. For example:
In the proposed font, if final yeh is to have an initial bump, it would not be the previous character.
Or drop the initial bump altogether.
alif would have a "serif" at the bottom (only) to distinguish it from Latin lower case l.
This matches Hijazi script.
Again a personal preference. Perhaps meem's eye should be split by the horizontal joiner, whereas fa/qaf's is above.
