9000 » from archive
«LexHub - Data Sets» (might be interesting for computational linguists and the like among us) - http://lexhub.org/data_sets.html