@figuerres: I don't know, maybe they think that using a text-to-speech reader to read the text back to speech recognition programs to writing the sample in your code is doable? *joke*