Open source, open APIs touted as news tools

Open source, open "APIs" and the open scraping of new resources to create a custom "TV newscast" were among topics covered on Saturday at the Georgia Tech Symposium on Computation and Journalism.

Jacob Kaplan-Moss, lead developer at the Lawrence [Kan.] Journal World, made a plea for news organizations to use so-called "open-source" software for their backshop operations. He said this empowers creative web and database development by reporters and technologists and makes innovation occur faster. Kaplan-Moss is one of the key developers of the programming interface Django. He also cited Python, Drupal, Joomla and Ruby on Rails among other open-source programming languages or interfaces that may be useful to news organizations.

"In my idea future, this open-source future, there is no difference between writing stories and writing code. One is in English, one is in Python," said Kaplan-Moss, adding: "In the last couple of years this stuff has gotten a lot easier and more accessible to journalists, and if it isn't let me know about it and we'll fix it." Kaplan-Moss said news-organization information-technology executives often resist open-source and outsource software implementations.

Kristian Hammond, of the computer-science department at Northwestern University, offered a demonstration of NewsAtSeven.com, a website which allows a computer user to customize for personal viewing a TV-style newscast, complete with avatars speaking the news. NewsAtSeven uses information scraped from sources like Yahoo News and WikiPedia, and various video-news sources. Hammond's team has created a Facebook application which allows a Facebook user to push a personally-created video into the NewsAtSeven service, where Facebook friends can embed it in their own newscast. Hammond said this raised the possibility of consumer syndication of user-generated news. When a user creates a virtual custom newscast, the URL for the cached newscast can be sent to friends.

All of the information pulled into NewsAtSeven is gathered using APIs, or application programming interfaces, available from the source websites, said Hammond.  One challenge NewsatSeven raises, said Hammon: What happens when a consumer pulls a news account from a source such as CNN's iReporter, which is video supplied by non-professional journalists. "You end up with a broadcast news show that no professional reporter was ever involved in," said Hammond. "It's an interesting thing to think about." In response to a question, Hammond said NewsAtSeven developers have not focused on copyright issues associated with scraping news from other websites without specific approval. He said they were still a research-oriented activity, subject to "fair use" copyright exemptions. He said most of the clips they pull are short.

John Brothers, CTO of the Sunlight Foundation, briefed the audience on the use of APIs to access government information on such things as legislative "earmarks", federal contracts, campaign finance records and legislation. He demonstrated Sunlight's offering a single line of Javascript code which allows websites to embed "popups" of politician mini-profiles as a link within text. Brothers demonstrated a partnership with MinnPost.com, in which MinnPost links photos of the Minnesota congressional delegation to detailed contributor and political information served up from Sunlight based on extensive mashing of API-obtained public data. "This is what we feel is one of the most powerful examples of using an APi to further the cause of demoracy and understand what's going on in our congress and the world," said Brothers.