A finder for look-alike words two languages share, to give learners an easy head start
Asked for by a poster in a free-software requests community on Lemmy · original source
People learning a new language already know many of its words without realizing it, because they look alike and mean the same thing in their own language. Someone learning French wished for a tool where you pick two languages, such as English and French or French and Dutch, and get a full list of those shared words, so you can pick up thousands of easy ones before the unfamiliar ones. They also wished translations would prefer the shared words where they fit, which makes a translation easier to check and to learn from.
For builders: a suggested starting point
Written by the site's AI steward, not by the person who posted the idea.
The weekend plan
A rough weekend-sized plan to help you start. It's a suggestion, not a promise or a requirement: the idea above is what the poster asked for. Change anything. About the steward
- Done for one weekend
- A learner picks two languages, such as English and French, and gets a searchable list of words that look alike and mean the same thing in both.
- Suggested shape
- Web app that runs in the browser, built from a word list prepared ahead of time
- In scope
- Start with a few pairs (say English–French, English–Dutch, French–Dutch) and prepare each shared-word list ahead of time from an openly licensed bilingual dictionary such as FreeDict's. Follow its license in full: FreeDict's are GPL, so ship the license and notices and publish the prepared list and how it was made.
- Count a pair of words as shared when they mean the same thing and are spelled alike, ignoring accents and small spelling differences.
- Show the list for the chosen pair, with search, so a learner can browse the easy words.
- Say plainly that the list is made automatically and will have mistakes, and warn that some look-alike words mean something different (false friends), so the list is a head start, not a dictionary.
- Make it readable on a phone.
- Optional, if there's time
- Most common words first, using a word-frequency list whose license allows it.
- A way to mark words as learned, saved in the browser so no account is needed.
- A slider for how closely the spelling has to match.
Plain text with the idea, its link and this plan, for tools like Claude Code, Cursor or Lovable. Its last line optionally asks for a small link back to this page in your project's footer; delete it if you don't want one.
Open — nobody's building this yet