Simplification proposal for the language file

[slaweallx]

Friends, Isn't it more logical to use an id value to make the language file simpler? Currently, the size of the language file looks extremely large and complex.

Let's give 6 digit random ID values now and in the future it will be easy for us to build a simpler and simpler structure. what do you think?

I would like to be personally involved in this task if it is accepted.

[titanism]

@slaweallx Hi there - we probably should switch to use LLM for translating these Markdown files and phrases instead of using Google Translate. Our current setup uses Mandarin, which breaks apart Markdown files and phrases into chunks, ignoring sections with notranslate HTML class names, and a bit more. If you'd like to dig in, look at @ladjs/mandarin at https://github.com/ladjs/mandarin, the files at https://github.com/forwardemail/forwardemail.net/blob/master/jobs/translate-markdown.js and https://github.com/forwardemail/forwardemail.net/blob/master/jobs/translate-phrases.js (which can be run one-off e.g. node jobs/translate-phrases.js currently).

[titanism]

Also it's in our build for Gulp at

forwardemail.net/gulpfile.js

Lines 672 to 685 in 89aedc2

	async function markdown() {
	const mandarin = new Mandarin({
	i18n,
	logger,
	...(isCI ? { redis: false } : {})
	});
	const graceful = new Graceful({
	...(isCI ? {} : { redisClients: [mandarin.redisClient] }),
	logger
	});
	graceful.listen();
	await mandarin.markdown();
	await graceful.stopRedisClients();
	}

[titanism]

We have a set of pre-defined phrases at config/phrases.js, but the Markdown files from index.md files in app/views folder get converted and parsed via Mandarin and broken into chunks, which is main reason why phrases are inaccurate in some languages and the reason why the locales/... JSON files are so lengthy.

[slaweallx]

My recommendation would be to use a language structure like the one I mentioned rather than automatic translation. I can do all the translations if necessary, but the structure we currently use makes this process extremely difficult.