Twitter sentiment analysis using Python and NLTK

This is a well written introduction to Bayes classifiers &#38; the NLTK API. It does leave out some important aspects, like tokenisation and testing. Perhaps add hyperlinks to further references, like the NLTK book (<a href="http://www.nltk.org/book" rel="nofollow">http://www.nltk.org/book</a>)?<p>Also, does anyone know where the 600/600 tweets comes from? It seems fairly arbitrary.