Re. Arabic text:

Do you plan to continue using LaTeX, or are you open to other software? I recommend taking a look at the Middle East version of Adobe InDesign, which includes the Tasmeem plug-in for which you can purchase special Arabic fonts including DecoType's Naskh and award-winning Nastaliq. In the context of your work, I think the great benefit of using Tasmeem is that your Arabic text would reflect the canons of Islamic aesthetics, i.e. the form of the text would reinforce what you are teaching.

For English text, you could do a lot worse than Arno Pro, and it will harmonise pretty nicely with a traditional naskh Arabic. Personally, I'm a huge fan of Kepler, which has a more modern feel than Arno, but it might not harmonise so well with the Arabic.

Perhaps the just released Latin-Arabic typeface family FF Amman may be of interest to you:

http://www.fontshop.com/fontlist/families/ff_amman/

PDF Specimen

Dear John,

I had Kepler in mind too, and for the moment I don't really need to write in Arabic. Do you think it's livelier than Arno for that kind of job?
I'll give InDesign a try in the future, but I'm afraid it will take me a lot of time learn the basics.

If he is going to teach Islamic art, then fonts like FF Amman are out of consideration.

Tasmeem is pretty good, but one need experience in Arabic calligraphy and rules, Tasmeem gives you tools to shoot yourself in the foot, I've seen both very good and very ugly (ugly to the degree that I'd consider it a calligraphic crime) work done with it.

Arabic typesetting looks best as a display font, but too light for body text (in my experience, people find it unreadable).

Dear Khaled,

I agree with you about Tasmeem. I realy think you need to be a very good Arabic calligrapher to profit from Tasmeem--which of course I'm not.

I was just looking to a readable font for small chunks of text in Arabic (instead of using only the Latin alphabet for transcriptions). If Arabic typesetting is not good for that, what alternatives are there? Scheherazade? Traditional Arabic?

I think Arabic Typesetting would do just fine for your use case, though you might want to set it in larger point size than the Latin text. You have to try your self anyway. Scheherazade have almost no ligatures except the four basic shapes for each letter, Traditional Arabic is better but still not as complete as it should (and some of its ligatures are rather weird or plain wrong).

Tasmeem comes with a set of Text Shaping pre-sets that can be relied on for basic styles of text, including one that represents a typical book style, the sort a scribe would use for copying a long text, as distinct from the kind of elaborations found in calligraphy. Yes, if one enters into the Word Shaping options or starts editing the Text Shaping options, then one needs to have a sense of the calligraphic aesthetics, but there's no need to engage in that side of Tasmeem at all: you can just select your text, switch to the Tasmeem Naskh font, and select ‘Naskh Standard’ from the Text Shaping pre-sets. You get something pretty typical of jobbing scribal work:

For modern texts, I prefer a more obvious distinction between inter-word spacing and intra-word spacing. These can be easily and globally edited using Tasmeem's Arabic Spacing controls.

John, foundry Arno lacks 02BE and 02BF. Of course, when you add them, they need kerning.

I'm not sure if d,h,s,t, and z with the dotbelow are in the foundry version, either.

Arno's native kerning doesn't survive a trip through Fontlab. If you want to add just one kern, you do all he kerning over.

From one whose been there and done it . . .

Charles, that's true. I hadn't considered the Latin font needs from the perspective of Arabic transliteration. That cuts down the options considerably, unfortunately.

The new version of Cambria that ships with Windows 7 supports all the necessary diacritics, including smallcap support.

The version of Minion that ships within the Adobe Arabic fonts does also, although it should be noted that this weight of Minion is incompatible with the common versions, as it was generated to harmonise with the Arabic; it also lacks smallcap support.

Dear Charles and John, the dots below were one of the reasons I turned to Latex in my thesis. With a small hack, I could insert dots below any character:

I didn't know that Cambria had the diacritics, though.

The Tasmeem example looks very nice, and doesn't seem very tricky. I'll definitely give a try in the next few weeks. Thanks a lot for the comments.

Your "hack" (AKA "macro" or "definition" in TeX) is just fine for preparing material for students. But likely your professional carrier will require both journal articles and books. Many of the journals do have a standard style for submitting manuscripts, books less so. Point is, editors will have to be able to open & work with your files. Oft times, LaTex (or any TeX) won't be acceptable. MS Word is the standard text editor, with the characters in proper Unicode.

The other issue is color management. Plain Tex, which we used with an extensive set of macros, wasn't good at it. This was the biggest reason we shifted to InDesign. InDesign's H&J algorithms are dreadful to one use to TeX, but while the color management can be difficult, it can provide what you need. Just how involved an author needs to be with color management is a separate question, but art usually involves images, and color management is critical there.

I used ConTeXt (a TeX based macro package akin to LaTeX and plain TeX) to typeset books that went for print, it had all colour management feature that I needed (my books are Arabic, with no single mathematical formula; not the typical TeX user). I know many journals that do accept TeX and LaTeX (and I know for at least one big scientific publishing house here that is based on LaTeX), but it depends on what field you are into.

When you are dealing with art -- esp. paintings, it is hard for those downstream to know what the original image really looked like. You can have all the technical capability in the world at the publishing/printing end, but even though you may know the characteristics of the press and paper the image will print on (i.e, have a perfect ICC profile for the press), the person preparing the art for printing will have no idea what the original looks like beyond the supplied image.

But you're right, the color management problems we encountered were in preparing mixed halftones and color, where the images ran with the text (not a gallery), using our version of Plain TeX. That may not hold for other implementations of TeX.

I really used Latex just for having a decent output with little effort. As far as I know, either no Middle East Studies or Arts journal would accept Latex.

I was just concerned to have a decent, nice-looking font (I still haven't decided between Arno and Kepler) and a good typesetting engine to translate some book chapters and papers for my students.

All printing will be done on a4 or letter paper, by a standard laser color printer (which I'll probably have to buy myself). All this work is because I fear very few students will be able to read the original textbooks, and as far as I know Brazilian publishing houses have no interest in publishing such time-consuming, fancy books.