corpus: data used for training

LM: language model built from training data

demo: demo used for skunkworksdata used for training

working: home of translation engine

translate: scripts to generate test data and translations