Hopefully they'll help clean up Wiktionary. I downloaded the English dataset the other day and it's a mess. IIRC, less than 1/3 of the words are properly tagged

Hopefully they'll help clean up Wiktionary. I downloaded the English dataset the other day and it's a mess. IIRC, less than 1/3 of the words are properly tagged as English words, and a significant fraction of the words aren't even rendered in the English character set. The good news is that the editorial standards are extremely consistent, so it's easy to extract information. They also allow batch updates, so if I find an easy way to correct vast swaths of this data, I could send a batch update file once