Looking at #Arabic#manuscripts with my students today. This is a stunning frontispiece from @subugoe Cod MS arab. 190, containing the Kitāb al-wāfīya fi šarḥ al-kāfīya from 1478, collated with the author’s own copy. @historikerinnen@medievodons@histodons#palaeography
Looks like today I finally found a good application for #llm 's: Learning languages!
I've been attempting to learn #arabic through duolingo for a while now, without much success. I figured if there's one thing language models should be good at, it's languages. So far the thing has actually been pretty helpful.
Turns out rendering #Arabic is hard. No, you can't just implement the #Unicode#Bidirectional Algorithm and call it a day. It turns out the Arabic letters/symbols/? have different forms depending on where they are in the word and probably there are other non-trivial features. Yeah, I guess I just postpone this.
So my 1000 IQ workaround for now is to just render all arabic characters as U+FFFD REPLACEMENT CHARACTER for now.
@liaizon as an #Arabic speaker, if I "read/hear" Al Shabab in a non-arabic context, I'd associated it with #Somalia islamist group. If it's Arabic context, then it's completely different. I guess the same when interacting with words such as Taliban (students?), Mullah (religious leader/teacher?), or ikhwan (brothers)
Here's a #linguistic fun fact, the #Arabic word for constitution دستور is borrowed from #Persian. And the Persian word for constitution قانون اساسي is borrowed from Arabic.
On our most recent visit to the family in Egypt I just became painfully aware that I’m the only person on the family who doesn’t speak Arabic (though our youngest child only speaks fairly little Arabic he understands a lot more, and the older one speaks it fluently), and that’s just really annoying.
All translations of literature seem to contain unfortunate mistakes that will probably never be corrected. Here in Leri Price’s translation of Khaled Khalifa’s novel No One Prayed Over Their Graves (a good translation overall), الفتيات السافرات means girls who don’t wear the hijab, not girls who travel. #literature#Arabic#translation
The Ruins of the Future: Speculative Fiction in Arabic
"#SOAS shifted the conversation around Arabic #scifi from lamenting the supposed lack of #Arabic texts in the genre to exploring nuanced ideas of dystopia and alternate temporalities."
...not only, but also the #Inquisition destroyed so much knowledge, which partially/not rarely survived in #Arabic translations , so we know about it again nowadays.
Also, to be fair to him, I'd need to read his book, as we all tend to simplify things orally/in interviews.
I were at #FOSDEM to show and discuss #fedora#languages#translation progress, and instead I went to #OFFDEM in which I had a really pleasant time, unfortunately, I were not able to stay on Sunday for personal reasons...
I'll try talking here about these data, until I consolidate this in a more structured way.
I a reminder, my goal is to measure the efficiency of the way we work (translator communities, devs and distribution), to open discussion.
Here is an example graph for #arabic language. Scope: #fedora operating system.
Is this Arabic translator community healthy? No
Why ? because the overall translation progress is decreasing over time, despite the number words increasing, it means there is more work than what translators can provide.
That's a pity, Wikipedia tells me there is 360 million speakers of this language! #translation#OpenSource
Just learned the #Spanish phrase "al azar" (by chance, at random) and found it is related to the #Bulgarian word "зар" [zar] and the #Greek "ζάρι" [zári] meaning "a die" (singular of dice), coming from a medieval #Arabic word for "a die" - az-zahr 🎲
Apparently the #English "hazard" is also a relation!
Tonight I started a discovery course of Hebrew. Just for me
I checked the alphabet
Similar system of consonants and strong vowels on which you append unwritten short vowels
Then I went on with the words that are close to Arabic, which I started to learn 20 years ago (not seriously enough): I still remember though how to pronounce most of arabic letters, and remember some vocabulary
I post my "course", because I recognize many of the words: THREAD will be LONG