Technology

Chuck Knol

Why Google’s online encyclopedia will never be as good as Wikipedia.

There are two articles about Sarah Palin on Google Knol, the search company’s abysmal new Wikipedia-like reference guide. One of them is a mess: Just a few hundred words long, the article is fraught with factual and grammatical errors. The other Palin entry is much more readable and informative, offering a thorough, balanced look at Palin’s years in city and state government and her positions on national political issues. Unlike Wikipedia, Knol displays its authors’ names and credentials to help you decide whether to trust a given piece. When I click on the name of the second Palin entry’s author, Sam Goldfarb, I see that he’s also written Knol articles about advertising on Facebook, the Chinese territory of Macau, and several hotels in Israel. How does Goldfarb know so much about so many things? You might call him a keen student of the Web—a bit of Googling confirms that each of his articles was lifted from other online sources.

Goldfarb’s great Palin entry is a copy of the Wikipedia article on the Alaska governor as it appeared on Aug. 29, the day John McCain picked Palin as his running mate. That’s why the Knol piece still describes Palin as having “successfully killed the Bridge to Nowhere”; the Wikipedia entry on Palin has since been updated thousands of times, and it now tells a more nuanced story about her flip-flop on the bridge. (Wikipedia’s articles are licensed under the GNU Free Documentation License, which allows people to copy an entry’s text as long as they also reproduce the license; Goldfarb’s Palin article and many others on Knol that copy from Wikipedia don’t follow those rules.) Goldfarb’s Macau article is lifted from this Macau travel site, his Facebook piece draws from this ad company, and his hotel guides pull from the hotels’ Web sites.

Knol is a wasteland of such articles: text copied from elsewhere, outdated entries abandoned by their creators, self-promotion, spam, and a great many old college papers that people have dug up from their files. Part of Knol’s problem is its novelty. Google opened the system for public contribution just a couple months ago, so it’s unreasonable to expect too much of it at the moment; Wikipedia took years to attract the sort of contributors and editors who’ve made it the amazing resource it is now.

But Google has grand ambitions for Knol. In a December blog post announcing the project, the company’s engineering chief, Udi Manber, wrote that Google wants Knol articles to stand as “the first thing someone who searches for this topic for the first time will want to read.” Of course, the first thing you find for many topics you search for now—including Sarah Palin—is a Wikipedia article. Unless Google radically redesigns Knol, it looks unlikely to supplant Wikipedia. The project suffers two critical flaws that promote poorly written, poorly sourced, and plagiarized articles. First, Knol diminishes community involvement, giving authors complete control over their postings. Second, it rewards authors with advertising lucre, creating a huge incentive for people to post as much content as possible. That probably helps explain why so much of Knol’s content is repurposed from elsewhere.

These aren’t haphazard mistakes. Google put these two measures in place by design to differentiate Knol from the world’s pre-eminent online encyclopedia. Wikipedia operates on a principle known as NPOV—contributors and editors aim for a flat, “unbiased” tone and a “neutral point of view.” Wikipedia is functionally anonymous. You judge the reliability of any Wikipedia piece not on the strength of the writing or the credentials of its authors but, instead, by the documents it cites to support its statements. (Never trust a Wikipedia article riddled with “citation needed” warnings.)

Google says neutrality and anonymity are overrated. Instead, Knol prizes personality and expertise. “We believe that knowing who wrote what will significantly help users make better use of Web content,” Manber wrote last year. While Google encourages authors to add citations in their articles, you’re supposed to judge a Knol piece not by its references but by the credentials of its author and the force with which he makes his case. That’s why Knol allows different people to post different articles on Sarah Palin. Competition between authors, Manber argued, produces better content.

Google’s argument fits in with the long history of writing and publishing. After all, we read books and magazines not for their neutrality but for an author’s clear point of view. Similarly, if you don’t know a thing about Roger Federer, you’ll learn much more from David Foster Wallace’s appreciation of the star athlete than from the Wikipedia entry that states in bland, NPOV language that “tennis critics, legendary players, and current players consider him the greatest tennis player ever.” So what’s wrong with encouraging a bounty of such articles online—a reference guide that’s both informative and stylishly written?

What’s wrong is that perspective and style don’t scale. Writing is hard even for the world’s greatest wordsmiths; it requires time, thought, and care. Good writing also usually requires good editing. Because Wikipedia’s NPOV guidelines set clear rules for what’s allowed on the site, Wikipedia is easy to edit—anyone can look up the tenets of NPOV and then set about cleaning up contributions that stray from the preferred style.

By default, Knol articles can be edited by readers, but each edit must be accepted by the original author before the revision takes hold. Along with the obvious problem of giving authors control of when they’re edited, Knol doesn’t give readers any guidelines for how to edit. One Knol article on Tori Amos describes her 2007 album American Doll Posse as marking a return to “daring and somewhat angry” songs and adds that her voice on the record sounded better than it has “since 2001.” Those lines are vague and mushy: What about the album is angry? Why does her voice sound so much better than before? Under Wikipedia’s NPOV rules, both descriptions would have to go, and any reader could delete them with a couple of keystrokes. But Knol allows such personal opinions, so you’d have to persuade the writer to excise them on other grounds—which, of course, takes a lot of work. Instead of going through the trouble, I clicked away.

As I perused Knol over the past couple of weeks, I tried to contact the authors of the few articles that I found interesting. This proved difficult; Knol doesn’t require writers to post their contact information. Even though readers are asked to accept these people as experts on a topic, there’s no easy way to ask them questions about their expertise. Still, I did manage to contact a few Knol posters, and I was surprised by what I found: Most people who contributed to Knol did so for money.

Some authors wanted to test the power of Ad Sense, the text ads that Google lets writers place alongside their articles—Google gives authors a share of the revenue it generates from those ads. (People told me they hadn’t earned more than a couple of dollars from these.)

Other authors were interested in promoting their Web businesses. Natasha Derrick, who runs a Web site called Hawaii Travel Guide with her husband, had repurposed several of her old pieces on Hawaii to post on Knol. (The Knol pieces include links to her Hawaii Travel Guide.) Derrick told me she hasn’t seen any increase in traffic to her site yet. But she and her husband see Knol as a way to “get in on the ground floor” of something great. Knol could be the next Wikipedia, and Derrick’s piece on the Hana Highway might make its way up the search rankings, delivering throngs of people to her site.

Derrick’s plan dovetails nicely with Google’s: If Knol is the next Wikipedia, both the writer and the company make a killing. The problem is that we don’t need the next Wikipedia. Today’s version works amazingly well.