2. DoenetML
3. DoenetML
4. Content
5. Bambara

Welcome

Thanks for helping translate DoenetML, an open-source platform for interactive educational content.

Every non-English string here is currently an unreviewed machine translation. None has been checked by a fluent speaker. You don't need permission or coordination to fix one — if it's wrong, just correct it.

What you're translating

Strings are split into four files (components in Weblate) per locale:

• chrome — viewer UI: buttons, labels, feedback headers

• content — worker-generated document prose

• diagnostics — warnings and error messages authors see while editing

• editor — editor and language-server UI

  English (en) is the source of truth. New keys always originate there — never add a key that doesn't already exist in English.

Placeholders and variables

Many strings contain Fluent variables like { $name } or { $count }. Keep these exactly as they appear — same name, same $ prefix, same braces — just move them to wherever they naturally fall in your language's word order. Weblate's placeholder checks will flag it if one goes missing or gets mistyped.

Some strings use Fluent's plural selectors ({ $count -> [one] ... *[other] ... }). Translate each branch; add or drop branches as your language's plural rules require (Weblate shows you the right branch set for your locale automatically).

Terms and attributes

Strings occasionally reference attributes on another message. Leave term/attribute names untouched — only translate the text they resolve to, in that term's own message.

Diagnostic codes

Diagnostic messages are tied to stable codes shown alongside the text (e.g. doenet-e0002). Translate the message text; never alter the code itself.

What not to translate

• Do not translate or touch the en-XA / en-XB pseudo-locales — these are synthetic QA locales used to catch layout and hardcoded-string bugs, not real languages.
• Leave any string you're unsure about in English rather than guessing — an English fallback is always safe; a wrong translation isn't.

Review and merge

Weblate commits translations to a weblate branch and opens a pull request against main in the normal GitHub flow, so your changes go through the same CI checks (syntax, placeholder parity) as any other contribution before they ship.

Questions? Open a discussion on the DoenetML GitHub repo or ask in our community discussions.