Abstract

Using an 11-million word corpus of the best works of twentieth century science fiction, we provide a thick description of an algorithm that generates a new, representative work. This chapter traces, through code, the successive formulations of the algorithmically generated text and proposes a method for analyzing algorithms and code.