20 comments

[ 2.3 ms ] story [ 54.4 ms ] thread
[I guess it's not your project. Some ideas anyway.]

It would be nice to make an online checker, where anyone can upload an image and the site tries to check if it's good (or at least not too bad).

I believe Google Translate can do that
Honestly if you get the 3 big things the link mentions you will be ahead of many things. Recognizing when the letters aren’t joined is (I think??) very doable even for people with no reading ability. Compare these two series of exactly the same letters:

کمی فارسی بلدید

ک م ی ف ا ر س ی ن ل د ی د

The tell for which one is wrong is the one where literally none of the letters connect. This is more common than you might think.

Well the examples looked a lot more similar in the text box before I posted them. RIP.
Also see

https://www.theguardian.com/tv-and-radio/2015/oct/15/homelan...

“In their eyes, Arabic script is merely a supplementary visual that completes the horror-fantasy of the Middle East, a poster image dehumanising an entire region to human-less figures in black burkas and moreover, this season, to refugees.”

To be fair, this is how pretty much any foreign language is treated in media, not some special treatment of the middle east in particular. You're lucky if they actually used google translate, and didn't just have the actor improvise.

Another big thing a lot of people don't realize is that google translate is really awful at some languages (especially those that are not related, and lack a lot of source text). An especially comical one is Latin, you see a lot of slogans and whatever translated to the language using GT, with 'Romanes Eunt Domus'-tier results proudly emblazoned on their corporate headquarters.

At my job in Middle East, my British Boss gave me a pdf poster about having some dimensions labeled in INCHES, but in Hindi language (the main language of labour on the ground). I asked him why this word (his google translation for inches) means IN in English? What happened, he input IN in the box, assuming IN as inches, where as IN means like in the water, in the box etc. Hindi doesn't have exact word for INCHES, its always just literal transliteration. My boss said, why would google do this, everybody knows INCHES are written as IN.
I have a ladder in my garage, with safety notices in English and Spanish. The English is dated May 20xy, the Spanish is dated Puede 20xy.
I had to deal with Arabic one for a coworker, in a database where most other data were in English. The main pain point was the right to left writing and thus the weird behavior for selecting text that occurred in the text editor when mixing scripts. I wish I could force selection to always be right to left even if that doesn't make sense in the target language but would make copy pasting easier.
I've seen the same with Hebrew. It's not nearly as widely spoken as Arabic but graphic artists like to include it on their designs and so on because of the unique letters which really contrast to the Latin alphabet.

Mistakes include: 1) Left to right letters 2) Bad translations, especially dating to the days when Google translate was not as good. 3) Incorrect grammar e.g. singular instead of plural, masculine words mixed with feminine. "Welcome" should be plural unless you really only want to welcome one person! 4) Spelling typos

If you are looking to learn Arabic but have always been scared by its perceived difficulty I cannot recommend language transfer enough:

https://www.languagetransfer.org/arabic

His courses are so amazing that I got interested in Greek after listening to a couple of his lessons and listened to the entire course when I had no prior interest before. It's that good.

Thanks for the recommendation. I have been doing Arabic on Duolingo since the start of the pandemic out of curiosity but find it sometimes lacking.
> We even have a website that will break our Arabic so that Photoshop breaks it back to normal. Yes, Adobe is aware. No, they do not care.
Arabic-script typography on the web can also be complicated. I did a review/demonstration of some fonts and issues that they may need to handle:

https://www.theobeers.com/arabic-font-tests/

The tl;dr for people who want Arabic-script text to look ok in their projects, without making it a focus, is that a modern "native font stack" (à la Bootstrap) does a fine job. I just wouldn't recommend allowing things to fall straight back to "serif" or "sans-serif."

I work day-to-day with Arabic and English and it often feels like Arabic web support is an after thought.

It feels like a 3rd or 4th class citizen.

I'm talking VSCode, Sublime, Safari, Chrome, XSL, Numbers, Word etc not just lazy content creation mentioned in the link.

Can only imagine how bad it must be for other writing modes.

Unless return from the extra investment to make the software Arabic-compatible is worth it for the company, nobody’s going to care.
This is a pretty nice summary. Could use some pointers on what to do next. Like making sure the characters are not printed from last to first when the letters are not properly connected. Or using dir=rtl if it's aligned wrong.