Information Theory provoked the interest of arts researchers from its inception in the mid-20th century, but failed to produce the expected impact partly because the data and computing systems required were not available. With the modern availability of data from public collections and sophisticated software, there is renewed interest in Information Theory. Successful application in the analysis of music implies potential success in other artforms also. I give an illustrative example, applying the information-theoretic similarity measure Normalized Compression Distance with the aim of ranking paintings in a large collection by their conventionality.