multiple lang translation aspects

[gedw99]

https://github.com/bouncepaw/melanocarpa/blob/main/css.myco is an example of a file with Russian and English in the one file.

I would like to be able to keep ( or at least know) what content is in sync with other content, in terms of language.

there are many options to help do this

1. Formal file structure:

"there was structure in docs wiki(? not sure where was it) like"
/hypha/very-helpful
/hypha/very-helpful/ru
/hypha/very-helpful/es
or
/hypha/en/very-helpful
/hypha/ru/very-helpful
/hypha/es/very-helpful

2. No file structure

Like this: https://github.com/bouncepaw/melanocarpa/blob/main/css.myco

One way we can do this to to add annotations to the file that tells the engine what language content is in:
"lang:ru" could be added to the css.myco file to tell the engine that its Russian, and the "lang:en" for the English content in the css.myco file. This allows any fiel structure to be used which might be preferred by some.

The other way is to detect it at runtime which i really dont like: https://github.com/pemistahl/lingua-go#library-reason, because we simply don't need it because we know when a user is adding content what language it is in, because we detect it at the web level of the architecture. Then again, I could be using EN locale and wanted to edit Spanish text :) It might be better then to do the locale check and assume its correct and then in the web gui, let the user override. It woudl only be for the "Create". It would not be for the "Change" of existing content.

then engine could then be altered to parse the annotation..

then a flow from Web can be:

• what content is missing / matching.
• anything else we might want to do... ...
• sent a notification that the EN version needs to be translated.
• do nothing if you don't really care .

The point is that it allows others to extend the system with different language content flows.

• spelling checking, thesaurus and other thing later too.
• cross language sync
• machine and human translation loops.

This is really just to at least have the ability to know the language that the content is written into for now.

Please let me know your own thoughts or ask me if there is not enough here...

[bouncepaw]

there was structure in docs wiki(? not sure where was it) like

https://mycorrhiza.wiki used to be the docs wiki. Both Russian and English were there, but Russian was deleted, because no one cared. It is not a docs wiki now. Docs are now built-in in the engine, i/e available on every single wiki. URLs of the documentation follow a pattern:

• https://mycorrhiza.wiki/help/en/hypha
• https://mycorrhiza.wiki/help/en/lock

As you can see, the language is part of the path to the document. We only have one language now for the documentation. We had both Russian and English, but Russian was deleted, because no one cared.

One way we can do this to to add annotations to the file that tells the engine what language content is in

What if we have multiple annotations in one file? What if I just put an emoji with Russian flag, doesn't it count as an informal annotation already?

we know when a user is adding content what language it is in

How do we know what language the user uses? Mycorrhiza accepts any text, they can write anything.

detect it at the web level of the architecture

We detect what language is preferred for the user interface.

then a flow from Web can be:

• what content is missing / matching.
• anything else we might want to do... ...
• sent a notification that the EN version needs to be translated.
• do nothing if you don't really care .

Re: notifications. We don't have them. Re: matching content. How would that work with annotations?

spelling checking, thesaurus and other thing later too.

That's a job for an operating system.

machine and human translation loops.

What is that?

This is really just to at least have the ability to know the language that the content is written into for now.

I believe people can distinguish languages very well without Mycorrhiza's aid. If someone cannot identify a language, then they don't care about that language.

Please let me know your own thoughts or ask me if there is not enough here...

Overall, I don't see what problem you are solving.