Wikipedia: Fact and friction By Richard Waters Published: January 1 2010 22:17 | Last updated: January 1 2010 22:17
It is one of the great paradoxes of the internet: an essentially open medium has created deeply entrenched information powers. Within a remarkably short time, companies such as Google, YouTube and Facebook have come to dominate the business of search, user-generated video and social networking – even though barriers to entry in theory remain low. Among the members of this emerging online oligopoly, none is more intriguing than Wikipedia. An outlandish experiment in communitarian action when it was launched in 2001, the user-generated encyclopaedia has become the world’s fifth most visited website and the first reference source for schoolchildren and adults alike – even if most swear a half-hearted oath to double-check anything they read there. It is now, in the words of Jimmy Wales, co-founder and guiding spirit, “part of the infrastructure of the internet”. And its influence is spreading: Mr Wales notes proudly that a version in Wolof (a language spoken by 40 per cent of the inhabitants of Senegal, according to its own Wikipedia entry) now has more than 1,000 entries. Yet in spite – or perhaps because – of its success, Wikipedia is proving slow to adapt. Even relatively simple ideas aimed at reinforcing its basic quality and reliability have turned out to be hard to push through. Take a plan to improve the English-language version, the most extensive with more than 3m articles, by subjecting changes proposed by newcomers to approval by more experienced editors and flagging up any revisions. A minor but highly symbolic innovation, this idea was introduced in the German version more than 18 months ago but its appearance in English has been delayed by disagreements over how it should work. A more limited set of changes is now set for early this year, Wikipedia says. “There’s an almost religious zeal among Wikipedians that everybody has an equal role to play in the system,” says Larry Sanger, who co-founded the site with Mr Wales but fell out with him over the direction of the project. Not that there seems much chance of a competitor displacing Wikipedia. As a charity that does not look to revenues, it hardly presents an attractive target for the internet’s other established powers. “If you look at Wikipedia as a business opportunity for Google, or anyone else, it’s a pretty terrible business,” says Mr Wales. “They don’t look at us and see a $1bn revenue opportunity they should be competing for.” Google’s attempt to create a Wikipedia-like service of its own, known as Knol, had only attracted 100,000 articles when the company last disclosed figures. “Not at all successful,” sniffs Mr Wales. With his service still less than a decade old and its evolution hard to predict, Wikipedia’s entrenched position raises uncomfortable questions. Wikipedians dispute recent claims by a Spanish academic, for example, that the ranks of active editors are declining, but concede that the site’s English version is undergoing an important transition. Sue Gardner, executive director of the Wikimedia Foundation, the charity that oversees the running of Wikipedia, says external data show that the editing ranks “started to stagnate” in 2006, though she adds that it is not clear yet what this portends for the service. Critics are more forthright. With editor numbers no longer growing in spite of an increase in articles, it will be harder to monitor quality – and vested interests could find it easier to make alterations that reflect their own views, warns Andrew Lih of the University of Southern California and author of The Wikipedia Revolution. “My biggest fear is a slow leakage of the truth, which might not easily be detected.” Any critique of Wikipedia’s fundamental quality has to begin with an acknowledgement: in the field of supposedly objective reference information, there are no absolutes. “How many publications in human history can you trust?” asks Craig Newmark, founder of Craigslist, the online classified advertising service, and a Wikimedia Foundation adviser. “All have been subject to some disinformation that has been put in there by the powerful; disinformation has been a normal part of human reality forever.” Idealists such as Mr Newmark argue that, freed from the inherent flaws of publications dominated by a narrow range of interests, Wikipedia could become “more reliable than anything we’ve ever seen”. Yet even optimists such as he agree with the more sceptical observers on this: that, in terms of reliability and service, Wikipedia still has a long way to go. That makes the foot-dragging over flagged revisions look like a bad omen. It points to a fundamental tension at the heart of Wikipedia that has stalled its development. Founded on an ideal of complete openness, any adjustment that seems to favour one group of contributors over another can seem like a betrayal of principle. “They’re sending out a message, even if in a quiet way, that some editors are more important than others,” says Matt Curinga, a software programmer and occasional contributor who is working on a doctorate on services such as Wikipedia. While technical considerations have contributed to the delay, “the majority of people didn’t like the idea”, adds Mr Lih. “I’m certainly not sure that the English-language Wikipedia, with its free-wheeling culture, is ready for that yet.” If even a modest idea such as this – aimed mainly at preventing vandalism to articles by casual visitors – can be stalled, it does not bode well for more important changes that will be needed in the future, adds Mr Sanger. There are two potential longer-term approaches to boosting the site’s accuracy and reliability. The first, favoured by the powers that be, involves refinements to mechanisms that enable self-policing. With the right processes, the interaction of the crowd will produce the best results, suggests Mr Wales. Quality “emerges from lots of participation”, he says. “It’s just giving the people who are participating the tools they need.” In future, adds Mr Newmark, it may be necessary to track the identity of editors, or find ways to measure their reputations, then rank their work – ideas that seem anathema in the present culture. When it comes to trying to distinguish good editors from bad, “we don’t have good, consistent measures, which is a problem for us”, says Ms Gardner. A related push involves helping readers to distinguish the reliable from the unreliable. The problem is not necessarily that the average quality of articles is low; rather, that there is no way to tell which can truly be relied on. Little effort has been put so far into analysing overall quality – a situation about which Mr Wales expresses frustration – let alone the quality of individual articles or the editors who compose or revise them. Nor do the search engines that generate so much of Wikipedia’s audience attempt to discriminate between reliable and less reliable articles. Some 60-70 per cent of the site’s traffic comes from search engines, says Mr Wales, and the fact that it appears at, or close to, the top of many search results is a big source of its success. One potential approach is to use internal clues in Wikipedia itself to make educated guesses about which information is most reliable. For example, Google could analyse the records of individual editors, says Marissa Mayer, Google’s head of search products. That information could be used to rank articles in its search index. Yet she says Google has no plans to do that, which means reliable and suspect articles alike will continue to figure prominently in search results. Wikipedia is testing its own tools for helping readers assess quality. For example, an experiment at the University of California at Santa Cruz colour-codes different segments of text in articles based on how long they have remained unedited, on the assumption that the most reliable parts remain untouched the longest. Wikipedia also plans this year to add a “rate this article” option to help guide readers. The second potential approach to boosting quality is more fundamental and raises deeper questions about the direction of the site. “What they really need is some sort of rating or approval process that engages actual experts,” says Mr Sanger. The idea that experts should have a superior status is a taboo, guaranteed to provoke a strong backlash. For a site – Wikipedia prefers “community” – founded on ideas of radical egalitarianism, it smacks of revisionism. Yet even some enthusiastic supporters say there is no alternative in the long term. “You need experts balanced by citizens, and vice versa,” says Mr Newmark. Informally, Wikipedia is already trying to draw in more experts. To overcome the strong resistance of many in academia, Ms Gardner says the foundation has been encouraging academics to contribute more. She quotes with approval experiments in which professors have set their students projects to improve certain articles. Yet that may have a lasting impact only if the work of experts is clearly accorded a higher standing. “Ultimately, that’s what has to be done – I just don’t think they’ll do it,” says Mr Sanger. While this struggle over the soul of the internet’s most successful experiment in user participation rages, the watchword for visitors to Wikipedia will remain the same: caveat lector – or, as the site itself helpfully translates, “Let the reader beware”. ------------------------------------ -- INFORMATION OVERLOAD? Get all ZESTMedia mails sent out in a span of 24 hours in a single mail. Subscribe to the daily digest version by sending a blank mail to [email protected], OR, if you have a Yahoo! Id, change your settings at http://groups.yahoo.com/group/ZESTMedia/join/ PARTICIPATE Share media news, discuss journalism issues and network with media professionals across South Asia on this mailing list. Just write to [email protected] TELL FRIENDS TO SIGN UP If you got this mail as a forward, subscribe to ZESTMedia by sending a blank mail to [email protected] OR, if you have a Yahoo! ID, by visiting http://groups.yahoo.com/group/ZESTMedia/join/ Also have a look at our sister list, ZESTCaste: http://groups.yahoo.com/group/ZESTCaste/Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/ZESTMedia/ <*> Your email settings: Individual Email | Traditional <*> To change settings online go to: http://groups.yahoo.com/group/ZESTMedia/join (Yahoo! ID required) <*> To change settings via email: [email protected] [email protected] <*> To unsubscribe from this group, send an email to: [email protected] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
