This Twitter bot generates tweets using two data sets: the text of the MIT Press book 10 PRINT CHR$(205.5+RND(1)); : GOTO 10 and any tweets with the #10print hashtag. The generator uses a Markov chain process to analyze a text and determine the probability of any word following another to generate a new string of words that resembles the original and publishes it on Twitter. Here are some examples:
How do we describe these generated tweets? Are they sentences? Phrases? Lines of verse? Would a more generic term be more appropriate as a descriptor, such as “text string?” Perhaps. But what’s the fun in that? And it is a fun project— certifiably so, since it was the 2nd runner up in the “Best use of DH for Fun” category of the 2012 DH Awards. Tweets like the third or fourth quoted above create grammatical sentences that are illogical or false, but interestingly so. To think of a Maze War as a framework for studying the telephone book is provocatively absurd. And there are other gems to be discovered in this bot’s tweet stream.
Poetically, this could be described as Flarf, a next-generation cut-up, or a kind of Language practice. It is also a generative language maze: taking the logically organized lines of prose that the book is written in and twisting them into obliqueness, creating a complex emergent set of twisty little passages.
Featured in Genre: Bot
