Lettercove word lists: sources and licences
===========================================

The files in words/ were built for Testera's Lettercove game by
scripts/lettercove/build_word_lists.py, from two openly licensed sources.
They are adapted material and are shared under the same licence,
Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0):
https://creativecommons.org/licenses/by-sa/4.0/

1. Wiktionary (English edition), via the kaikki.org machine-readable
   extracts made with wiktextract by Tatu Ylonen.
   Used for: which spellings count as words, and the short English meanings.
   Licence: CC BY-SA 4.0 and GNU FDL, as Wiktionary's text.
   Source: https://kaikki.org/dictionary/ (French, Spanish, German, Irish,
   English, Indonesian), and https://en.wiktionary.org/ with its
   contributors.

2. FrequencyWords by Hermit Dave, from the OpenSubtitles 2018 corpus.
   Used for: how common each French, Spanish, German, English and
   Indonesian (Bahasa Indonesia) word is,
   which sets the learner-sized cut (the 50,000 most frequent forms) and
   what the bots and hints prefer.
   Licence: CC BY-SA 4.0 (the word lists; the repository's code is MIT).
   Source: https://github.com/hermitdave/FrequencyWords

Irish has no FrequencyWords list, so Irish words are ranked by how often they
appear in the Irish example sentences in Wiktionary itself.

Changes made: words were lower-cased (German nouns keep their capital in the
meanings files), folded to the game's tiles (French and Spanish accents
removed except Spanish ñ; German ß written as ss; Irish fadas kept;
Indonesian words with Q or X left out, as those letters have no tiles),
filtered to words of 2 to 13 letters, with proper names, abbreviations,
obsolete forms, words marked offensive or vulgar in Wiktionary and a short
hand-made list of swear words in each language removed,
and meanings shortened.

Lettercove's code, board, bonus pattern and tile sets are Testera's own
(Aqta Technologies Limited, Dublin).