Definitely, you on the right way 😉

Well, the Linguist is no have any public opinion. You are welcome to use the unique project who care about UX, and don’t care about political views of their users.

Sure, decision logs is just a step to fix a problems with decision making. Logs highlight a problems in decision making process and let you analyze a problems in your team. You still need an strong architect who may say “hey team, why your arguments so weak? it looks you are low skill engineers. make arguments more objective, or you will be fired” and then really fire weak engineers in team.

Exactly! I see how a mindset “we delivered (shit) quickly” successfully reproduces itself, over and over in a lot of companies.

I think the actual reason is that business want to make money on promises right here right now. Business promises a lot of features to investors give money, and then business implement a lot of (shit) features. And then they continue kinda “okay, features broken, give us more money and we will fix it”.

Haha, actually it’s my bad, not a Linguist 😀

I’m not a native speaker, i still learn the language, but you can edit a post text on a github https://github.com/vitonsky/blog/blob/master/posts/2023/july/linguist/2023-07-13-linguist.md i would appreciate it

deleted by creator

Yes, the links you needs:

• Linguist docs about custom translators https://github.com/translate-tools/linguist/blob/master/docs/CustomTranslator.md
• Code of custom translator for a LibreTranslate https://github.com/translate-tools/linguist-translators/blob/master/translators/LingvaTranslator.js
• Guide to deploy a LibreTranslate on local machine https://github.com/LibreTranslate/LibreTranslate#install-and-run (or to use with docker: https://github.com/LibreTranslate/LibreTranslate#can-i-use-libretranslate-behind-a-reverse-proxy-like-apache2-or-caddy)

A LibreTranslate may download a gigabytes data once, to get AI models for all language directions, then translation will runs on your computer

You can make a language request in a issues of repository https://github.com/mozilla/firefox-translations-models/

You’re right, resources are limited and a new languages appears not so fast as we wants. Maybe Chinese developers will collaborate and release their own models for a Bergamot. Work with community are important to achieve the goal when we want to unify translation.

If you really needs to be private for now, you can try to deploy your own offline translation server locally (e.g. with a https://github.com/LibreTranslate/LibreTranslate) and use this translator with Linguist as a custom translator.