Jacob Harris, a software architect at The New York Times, has created an algorithm that reduces articles on its Web site’s homepage to poetry — specifically the form commonly known as haiku. The code mines words from each day’s top stories and assembles them with the help of an open source pronunciation dictionary, and the results appear on a Tumblr titled Times Haiku that’s hosted by the company. The “haiku bot” and Tumblr launched yesterday, just in time for the start of National Poetry Month.
What’s the Big Idea?
A quick review of the Times Haiku site reveals poems that, in a strange way, draw readers to the underlying article (linked at the bottom of each entry). That’s the goal, says assistant editor for interactive news Marc Lavallee: “If someone sees the site, or the image of an individual haiku and shares it on Tumblr, and it…gives them a moment of pause, I think we’ve succeeded in a way.” Harris says that although the project is relatively small and may seem silly, the code underneath could be valuable for future efforts involving the Times’ “large corpus” of articles.
Charles Darwin and Richard Dawkins may never have envisioned the current era of Big Data, but their shared fundamental principle – “all life evolves by the differential survival of replicating entities” – […]