MotsDeMinuit

Attribution — licence des données

This is the official attribution and licensing statement for the third-party data MotsDeMinuit bundles and derives from. It applies to every release (FOSS APK, Google Play, F-Droid) and to any redistribution of the derived files.

1. Dictionary data — Wiktionnaire (French Wiktionary) & WikiWoordenboek (Dutch Wiktionary)

  • Sources: full XML dumps of the French Wiktionary, https://dumps.wiktionary.org/frwiktionary/latest/ (frwiktionary-latest-pages-articles-multistream.xml.bz2, retrieved 2026-08-31), and of the Dutch Wiktionary (WikiWoordenboek), https://dumps.wiktionary.org/nlwiktionary/latest/ (nlwiktionary-latest-pages-articles-multistream.xml.bz2, retrieved 2026-09-18).
  • Derived data: the word lists, lemmas, grammatical types, morphological tags and short definitions produced by the extractor in dictionnaire/ per language (documented in dictionnaire/tags.md): French under assets/dictionary/fr/ (mots.csv.gz + features index and definition shards), Dutch under assets/dictionary/nl/ (same three outputs).
  • Licence: all Wiktionnaire and WikiWoordenboek text is available under Creative Commons Attribution-ShareAlike 4.0 (CC BY-SA 4.0) — https://creativecommons.org/licenses/by-sa/4.0/deed.fr (dual-licensed with the GFDL; we rely on the CC BY-SA layer).
  • Compliance:
    • Attribution — the in-app « Licence des données » screen and the store listings credit the Wiktionnaire and the WikiWoordenboek (block below). The full revision history of every entry is available in the source dumps.
    • Share-alike — the derived files (mots.csv, its compressed forms and any filtered word list shipped in the app, French and Dutch) are made available under the same licence, CC BY-SA 4.0, published in this repository under assets/dictionary/ and in dictionnaire/ tooling, together with the extraction code that produced them.
    • Modification notice — the data was extracted and cleaned from the dump (wikitext parsing, deduplication, minimum word length, morphological tag extraction); these changes are shared here.
  • No media: no image, audio or pronunciation file from the Wiktionnaire is included; this statement covers text data only.

2. Usage frequencies — OpenSubtitles corpus

  • Sources: frequency lists fr_50k.txt (top 50,000 French tokens) and nl_50k.txt (top 50,000 Dutch tokens) from hermitdave/FrequencyWords, generated from the OpenSubtitles 2018 corpus (OPUS, https://opus.nlpl.eu/OpenSubtitles2018.php).
  • Licence: CC BY-SA 4.0 for the data (MIT for the repo’s code).
  • Obligations: same as above — attribution given here and in-app, and any frequency-annotated derivative file of this project (the normalized fr_frequency_50k.normalized.txt and nl_frequency_50k.normalized.txt assets) is distributed under CC BY-SA 4.0.
  • A language-resource alternative for frequencies is Lexique 3.83 (https://www.lexique.org, LGPL-LR); it may replace or complement this corpus later without changing this statement’s structure.

3. Relationship with the app’s own licence

The application code is GPL-3.0 (see LICENSE); these are data licences. A GPL-3.0 application may bundle and serve CC BY-SA 4.0 data; the data keeps its own licence and remains free.

This file must be kept in sync with the actual bundled assets. When a new dictionary snapshot is imported (see the regeneration instructions in dictionnaire/README.md), update the retrieval dates above.