Missed Connections (2007)

Missed Connections is a 2-Channel internet-aware software piece that continuously fetches the latest posts in the "missed connections" section of Craigslist.org. Each post is presented one at a time, and is filtered by looking for so-called stopwords. Computer Scientists define stopwords as those words that do not convey the meaning of a message. In essence, they are considered signal noise in the stream of potential information. Each post is presented simultaneously in two ways: one just with stopwords, the other with non-stopwords, and in both cases the filtered words are displayed as dashed lines, akin to the way words are presented in the game Hangman. Thus, both posts present the same "graphical" structure, but have the potential for very different readings.



The piece uses Craigslist's RSS feature to obtain new feeds to add to the XML database that the software uses. Once new feeds are obtained, a screen scraping routine is employed to obtain the full text of the post. The software operates in real time, but it keeps a cache of posts to cycle through. This cache is periodically flushed, its period determined by the number of missed connections posts that the program obtains in a day. Like many of my other pieces, Missed Connections was developed in Java. XML reading and writing was made possible via JDOM and the RSS component used ROME for parsing the feeds.