A Microformat with Major Implications

A vision of automated correction notifications and more

Imagine this: you visit one of your favorite news sites and the homepage displays a notification that an article you read yesterday has been updated with new information, and a story you read last week has been corrected. The notification enables you to click on a link and read the correction, or to be taken to the updated story.

After checking those items, you continue reading articles on the site, and each story includes a box of information explaining the type of sourcing used within the story (anonymous, etc.), as well as a link to the organization’s relevant policies and standards. If you spot an error in an article, you can easily submit a request for correction via that same info-box. And if the article is corrected, you’ll receive a notification during a future visit to the site.

This scenario is admittedly far from the reality of today’s online news experience, but we are closer to getting there thanks to hNews, a microformat for news being developed by the Media Standards Trust and the Web Science Research Initiative, of which Sir Tim Berners-Lee is a director. Their project won a Knight News Challenge grant last year; it also receives funding from the MacArthur Foundation. (Disclosure: as associate editor of PBS MediaShift, I also edit the articles that appear on MediaShift Idea Lab, a Web site that’s also funded by the Knight Foundation.)

The goal of hNews is to make certain elements of news articles machine-readable. In and of itself, this may not sound like much. One benefit is that it will help search engines and other parties do a better job sorting and analyzing information contained in news content (cover your ears, Rupert Murdoch!). In its current 0.1 release, hNews enables a creator to encode the byline, originating news organization, and date that an article was published into a story.

The average reader won’t see any of this behind-the-scenes action, but a news organization could use hNews to display this and other metadata in a format that becomes more user-friendly. For a rudimentary example, take a look at this article. Scroll down to the bottom of the text and place your mouse over the blue box labeled “Value Added.” (It’s located right above the comments section.) A small box of text will pop up that lists the article title, author, date published, and last date that the article was updated. It also includes a link to the Web site’s statement of principles. This information is automatically generated, thanks to the use of hNews.

The goal is to eventually include much more detailed information. But, in a larger sense, hNews is about bringing more transparency to news, according to Martin Moore, director of the Media Standards Trust. It will provide people with the back story and context of an article they’re reading.

“This is not the holy grail, it’s not going to suddenly enable people to get really high quality news, but it will help you get more transparent news,” he says. “It’s not going to suddenly mean everything encoded with hNews is triple-sourced and totally accurate. But it allows people to… judge news more easily because the information contained within is made more transparent and visible.”

Here’s how the hNews team described their vision when they received their Knight grant:

The plan: to design a way for content creators to add information on their sources to their reports, as a form of “source tagging.” For instance, a reporter could note that an article was based on personal observations, interviews with eyewitnesses or specific, original documents. Filters would then use this data - the “story behind the story” - to help find high-quality articles. A reader searching the phrase “Pakistan riots” for example, might find 9,000 articles. But filtering by “eyewitness accounts” would yield a more selective list.

As is my wont, I’ve taken their vision and molded it to fit my own corrections-oriented needs. This speaks to my personal interests, but it also helps communicate the potential value of hNews.

Before they can move to the next level, Moore says the team needs to see how its basic version will work in the real world. Right now, the Associated Press and AOL are encoding articles using hNews. (Read this recent blog post by Moore for more information about those relationships.) Eventually, the encoding will become richer and complex. And hopefully lots more news organizations will sign on.

“The first stage was getting really basic stuff accurate and marked up in a consistent way,” Moore told me. “The future aspects that are slightly different, and in many ways really interesting but obviously much more complex, are to do with the state of the article over a period of time, the sourcing of the article, or the particular people and places within the article.”

They also hope to enable a form of version history to track changes that have taken place within a story over a period of time, much the way Wikipedia does. If they can do that, they could of course embed correction and update information in the article. That’s where my vision of automated correction notifications begins to take shape.

For example, The New York Times asks people to register for a free account on their site. If the Times decided to encode every story using hNews, and the specification was advanced enough to include correction and updated information tags, then it would be possible to let readers sign up for notifications about the articles they read. The key is that everything would be automated: the hNews encoding would be built into the Times’s CMS, and users would (opt-in to) receive the relevant notifications. Suddenly, you can put relevant corrections and updates right in front of readers, and corrections become embedded in the relevant articles.

I don’t mean to make it sound easy. hNews is in its infancy and its effectiveness will be largely determined by its adoption rate. If it becomes a de facto standard for online news, and is built into major content management systems, then suddenly a world of things become possible. But if hNews is used by only a few news organizations, its value will be greatly diminished. Success requires Herculean effort by the hNews team, open-minded thinking by major news organizations and CMS developers, and the backing of major Web standards bodies.

But this is one microformat that’s worth some major effort.

Correction of the Week

“On November 5 we translated the name of Ed and Nancy Kienholz’s artwork at the National Gallery, The Hoerengracht, as ‘Gentlemen’s Canal’. This should have read ‘Whore’s Canal’. We apologise for the error.” – The Times (U.K.)

Update 2:10 PM, 11/13/2009: This article was updated after publication to note that hNews also receives funding from the MacArthur Foundation.

Has America ever needed a media watchdog more than now? Help us by joining CJR today.

Craig Silverman is the editor of RegretTheError.com and the author of Regret The Error: How Media Mistakes Pollute the Press and Imperil Free Speech. He is also the editorial director of OpenFile.ca and a columnist for the Toronto Star.