Closed data-man closed 2 years ago
Should be done on the QOnlineTranslator side.
It's not a translator: Tatoeba is a collection of sentences and translations.
. Isn't this out of scope?
You are right, thanks.
What is a translator? You want to know how a word or sentence in one language would look (sound) in another language. So Tatoeba is the perfect translator, as the words/sentences are proper and made by native speakers. A word/sentences not found? OK, an empty output.
API returns data in pages. On each page can be different set of languages. Also, we can't control, which language will be in output. I can't see how it will fit what Crow is do right now.
What is a translator?
Something like Google Translator (Lingva - frontend to Google), Yandex Translator, Bing Translator, LibreTranslate. Pre-trained models, which can return some result based on static data. What you suggest is dictionary, not translator. And this dictionary is community-controlled, which is not a bad thing, but without moderation it is chaos.
I can't see how we can display all of this in Crow. For example, user want to translate what I just write. There will be no result, because this sentences is not in database. Empty output will be confusing.
If you will suggest how it will look in Crow Translate, GUI-side, I think we can implement this, at least implementing an backend will be simple.
Thanks for the app!
API: Maybe Tatoeba-API-specification-2 (I'm not sure).