Man Or Machine?
By Neil Dodds
14 March, 2006
Billions of web pages, 75,000 new blogs every day and much of the old media world gradually moving over to the internet. There's a lot of information out there. The question is, how do you find interesting news? And how do you sort out the good stuff from the dross?
In one corner, human filters and editors. In the opposite corner, algorithmic news aggregators, who 'learn' which stories appeal to your tastes and create front pages or newsfeeds tailored to your requirements. Google News leads the field in the latter category, though as Wired reports, there's a chasing pack of algorithm-based news finders promising to seek out stories from the depths of the web, direct to busy users' desktops.
The human end includes news sites like Digg, where registered members recommend or give the thumbs down to breaking stories, allowing users to get an up-to-the-minute view of what stories their peers are recommending to them. While the stories that hit Digg's top spots are generally technology related, it isn't difficult to imagine communities of users with other interests (the environment, celebrity gossip, finance and share dealing) creating sites along the same concept.
But surely there's still room for sites put together by an individual, or small group of individuals? Look at The Drudge Report. 11 million visitors a day come to Drudge for the Matt Drudge's discerning take on the news: A taste for the sensationalist, right-leaning, a heady blend of entertainment, politics and exclusive breaking news reports from the man himself. Then there's the Huffington Post, the liberal news site we've already discussed in more detail today. As well as commentary from academics, politicans and celebrities, HuffPost points to news stories likely to be of interest to its readers. Like Drudge, it draws on sources as diverse as AP feeds to YouTube videos; like Drudge, it brings in millions of daily readers who want their news chosen by enlightened individuals - not necessarily flash mobs roaming from story to story, and certainly not an automated algorithm.
It's the same story on some big name blogs: Indeed, one of the biggest, Instapundit, made its name as a linker to interesting news articles, interspersed with comments from blogger Glenn Reynolds. Instapundit used to boast the headline "The Grand Central Station of the blogosphere" - though since Reynold's career as a pundit has taken off, he's dropped the banner. Still, a good 90 percent of posts on Instapundit today are short links to other commentators, accompanied by pithy comments or selected quotes.
But relying on Reynolds, Drudge or Huffington means you're going to miss out on a lot of potentially interesting news stories. Similarly, setting up your own tailored RSS feeds based on favourite sources can be time consuming and equally limiting - especially if you don't read half the feeds you receive each day, as one contributor to the feature confesses.
Besides, RSS is not transparent technology for most users, even digital natives.
Wired suggests that machines could win the race, and speaks to a few companies fine-tuning their selection methods to make sure that every whim of potential readers can be catered for.
Fine - but every new filter adds another layer between the reader and the news: Wired's commentator speaks of an RSS feed pointing to stories your editor thinks you should be reading, which is then directed into a peer-reviewed site or another news editor, which filters again according to your tastes: What's distilled at the end of the process should be news worth reading.
"There's only so much I can read, really, and when I wake up in the morning I decide who my sources are for deciding what I should read," says tech guru Rael Dornfest.
If tech evangelists admit they can't be bothered reading the products of their finely-tuned algorithims, how can they expect ordinairy punters to do so?
Suddenly, paying a guy to tell you what's worth reading - in other words, buying a newspaper - seems like the most attractive option.
|
 Printer friendly version
 Email this article to a friend
Related articles:
|
