New language

[ghost]

I would like to add croatian. How can I do that?

[aradzie]

Hi!

I see you added a Croatian translation on POEditor. That's amazing! I have already updated the web site.

To add a training language, the biggest challenge is to find a word frequency dictionary, or a large corpus of text to scan for words.
I have developed my own set of tools for building word frequency dictionaries in a separate repository.

Over time I developed the following approach:

1. I scan a corpus of text built from movie subtitles for word frequencies. I chose this corpus because movie subtitles are written in a simple contemporary language, without a bias to a specific topic, like politics, news, etc. You can suggest another corpus, if you know a better one, like the Croatian National Corpus.
2. I send a list of words to a native speaker to review and remove obscene, profane, triggering words.
3. Then I include the reviewed list of words into keybr to add a new training langauge.

In the coming days I'll build a Croatian word list and send you for review.