John Allen Paulos in his column at ABC News:
The Ouvroir de Litterature Potentielle (Workshop of Potential Literature), Oulipo for short, was the name of a small group of primarily French writers, mathematicians and academics devoted to the exploration of mathematical and combinatorial techniques in literature. Founded in 1960 (and still somewhat active), the group searched for new literary structures via the imposition of unusual constraints.
Raymond Queneau’s “One Hundred Trillion Sonnets” is a prime example of Oulipo’s combinatorial approach to literature. The work is only 10 pages long with a sonnet on each page. Cut crosswise, the pages allow each of the 14 lines of each sonnet to be turned separately. Thus any of the 10 first lines may be combined with any of the 10 second lines, resulting in 10^2 or 100 different pairs of opening lines. Any of these 10^2 possibilities may be combined with any of the 10 third lines to yield 10^3 or 1,000 possible sets of three lines. Continuing, we conclude that there are 10^14 possible sonnets. Queneau claimed that they all made sense, although it’s safe to say that the claim will never be verified since there is more text in these 10^14 different sonnets than in all the rest of the world’s literature.
Another good example of Oulipo’s work is Jean Lescure’s (N+7) algorithm for transforming a text. Take an excerpt from your favorite newspaper, novel or holy book and replace each noun in it with the seventh unrelated noun following it in some standard dictionary. If the original text is well written, the resulting text is a bit surreal, but usually retains the original’s rhythm and even something of its sense. “Fourscore and seven yeast ago our fathoms brought forth on this continuance a new native, conceived in library and dedicated to the proprietor that all menageries are created equal.”
More here.