Hacker News Reader

Comment by antman | Hacker News Reader

Which languages are these? Especially the two that look similar

English, Assamese, Bengali, Hindi, Telugu.

The first, I expect requires no explanation!

The second and third are in the Bengali script, which is used primarily by Bengali and Assamese—but the letter ৰ is no longer used in Bengali.

The fourth is in the Devanagari script, which is used by many languages, but Hindi is the biggest one by far. (Second comes Marathi at about ¼ the native speakers, then Bhojpuri.)

The fifth is in the Telugu script which is primarily used by the Telugu language.

(Sources: personal knowledge of Bengali and Telugu and recognition of Devanagari, as an Australian who’s spent a lot of time in India, mostly in Telangana and West Bengal; and affirmation by Google Translate auto-detection of each one in turn, though that’s not very reliable as you’ll easily see if you try just pasting the whole lot in, and observe how it decides it’s Telugu, but still happily translates the others as “Merry Christmas”… except the Assamese, which it translates as “happy birthday”. Google Translate makes huge gaffes with Indic languages even if you identify the language correctly, and it just gets worse when it’s trying to translate it from a different language.)

It's technically the "Bengali–Assamese script". There is a difference of 3 letters between the written forms of these two languages, including the one you pointed out.

Google translate added Assamese recently, that might explain the gaffes.

If you want to be technical, you can call it Eastern Nagari. In practice, the name “Bengali–Assamese script” doesn’t seem to be used outside of Wikipedia and adjacent projects: it’s overwhelmingly called the Bengali script, with Assamese materials calling it the Assamese script. The fact that there are almost 20× as many speakers of Bengali as there are of Assamese probably contributes to this. (There’s also Manipuri, but it’s smaller still.) I find references to some ISO decision from June 2018 to recommend renaming the Bengali script to Bengali/Assamese script in Unicode (less than they were asking for, which was separate encoding due to some pronunciation differences), but I can’t see that this has happened. But I’m not familiar with the matter.

Google Translate has major gaffes in many of its less common languages. This is nothing particular about Assamese.

>I find references to some ISO decision from June 2018 to recommend renaming the Bengali script to Bengali/Assamese

Yes, that seems to be stuck and no progress made there AFAIK. For those interested, this Twitter thread covers the timeline - https://twitter.com/KabirFiraque/status/1066392176689643525

I just tried "শুভ বৰদিন" and Google translated it correctly as Assamese. Weird.

Bangla and… dunno the other one (Assamese?)

Reply on news.ycombinator.com