Quantifying the Dissimilarity of Texts