README file for Once Upon a Time / Monkeying Around I. DESCRIPTION It's hard to improve on the judges' description from DH2010 : 'Once upon a Time / Monkeying around tool was an original serious game that asks the user to write a story using the vocabulary of another author. The game provides a list of words the player must try to use in the story and then restricts the player to the vocabulary -- and the word combinations (bigrams) used by a particular author or corpus, such as Conan Doyle or Jonathan Swift.' See the video for a description of the theory and interface; basically, you are given a set of words from the authorial vocabulary and you try to use them in constructing a narrative _using only the author's style_ (as measured by bigrams. It's harder than it seems, and it appears to shed a lot of light on the idea of authorial style.... II. GETTING THE SOFTWARE 1) Download and compile the software; it's in a single .java file that uses basic Java/Swing. I've run it on a Mac (Snow Leopard) using Java 1.6, and it works fine. 2) Prepare a set of documents. For each authorial database (e.g. Foo), you will need to prepare three files. * Foo.start contains a simple string used as the "seed" for the story. Something like "once upon a time" (for a fairy tale) or "It was a dark and stormy night," or "I see by my records that it was the year 1885 when" (for a Sherlock Holmes parody). * Foo.1g contains all the "words" used in the document set. The words have been converted to all lower case and all punctuation has been stripped. * Foo.2g contains all the word bigrams (pairs) separated by colons. The program ships with two sets of data. One (ACD) from a collection of Sherlock Holmes stories by Arthur Conan Doyle (The Hound of the Baskervilles, the Memoirs of Sherlock Holmes, the Return of Sherlock Holmes, The Sign of the Four, and A Study in Scarlet) and the other (JS) from the Jonathan Swift Archive [normally available at http://www.arts-humanities.net/projects/swift_project] 3) Run the program using as an argument the basename (e.g. ACD, JS) of the dataset developed in the prior set. For example, to play with the Conan Doyle data (which is more fun), unpack the data files in a directory named "data" and then type java Monkey data/ACD or java Monkey data/JS for the Jonathan Swift data (which is frankly, less fun). 4) Have fun and ship suggestions to the author, Patrick Juola. I look forward to hearing everyone's feedback.