Comment by scott-smith_us

Is there significantly more to this than just tokenizing a document's words and then ZIP-ing the stream of tokens?