Thanks! Names are just chosen randomly from a relatively small vocabulary. The original idea was to query them from WordNet (or similar database), but that would be too much work :)
