Wikipedia's next steps

NEW YORK (Fortune) -- At my daughter's New York City high school her teachers tell her to not to use Wikipedia. When I talked to Jimmy Wales in Davos a few weeks ago, I expected him to be dismissive, perhaps even contemptuous, about such attitudes.

But he wasn't. Wales, who led the development of Wikipedia and who goes by "Jimbo Wales" online, is a nuanced, honest thinker refreshingly unlike the polished company-promoters I spend so much time listening to. "I always say it's a matter of using Wikipedia appropriately," he says. "I always laugh when somebody tells high school kids not to use it at all. That's like telling them not to listen to rock music. Whether it should be something you should cite in a paper, that's another issue."

So what constitutes appropriate use? "You should use it for broad background and some specifics. Like the World Cup of 1962. There's an article about it in Wikipedia, and it's almost sure to be reliable. But if the article talks about a goal being made at the last moment, that conceivably could be something a prankster slipped in and we missed. A lot of the errors in Wikipedia are ludicrous jokes that just get overlooked."

Wales showed me how you can check previous versions of a Wikipedia article by clicking on a "history" tab at the top of the page. You see the entire history of that article, including every change, and why it was made. Changes appear in red in the history versions. It's possible to compare any version of a story to any other. When we looked at the article about Tony Blair, he showed me how someone had put a line saying Blair once was a "supervisor at Toys R Us." But what's amazing is that the error was corrected a mere one minute after it was inserted. Next to the correction was a code indicating it had been vandalism.

Every change in Wikipedia is examined by software, what Wales calls "bots", which examine the change for incongruities. Changes that look suspicious are flagged in electronic messages that are sent out automatically to the thousands of volunteers who monitor specific articles. They quickly examine and change articles that have been vandalized. These changes frequently are done within minutes. Wales shows me another error introduced by vandalism. "That was there for 23 minutes, which was really slow," he says.

When people like my daughter's teachers assume that nothing in Wikipedia can be relied on because anyone can modify it, they fail to take into account the incredible ability of software to help police it. But Wales was also interested in talking about Wikia, the for-profit company he has co-founded to host wiki-style aggregations of information about anything. "Wikipedia is an encyclopedia," Wales explains. "The Wikia people are building the rest of the library."

There are 400,000 articles in Wikia so far, with especially deep information on pop culture topics like Star Trek, the game Doom, and TV shows like 24 and Lost. Complex TV shows generate huge numbers of articles about individual characters, episodes, plot lines, etc. There are 12,000 articles about Muppets at wikia, compared to 300 on Wikipedia, Wales notes. "The Muppet wiki might include an article on Itzhak Perlman," says Wales, "about the time he came on Sesame Street." Even "Uncyclopedia," a comic parody of Wikipedia, is hosted on Wikia.

But the part of all this wiki-empire that Wales is focusing his personal energies on most is what he calls "a completely open-source search engine." The idea is to use Wikia's infrastructure to develop a search tool for the Internet where the rankings of a given site are determined by a community of people who are interested, analogous to the volunteer community that oversees Wikipedia.

At the project homepage Wales asks for help in developing something "radically new." In our conversation, he explains that his desire for a new search engine is in part what he calls "political." "Search is a fundamental part of the infrastructure of the Internet," he says, "and unlike most of the infrastructure, today it's not transparent. We have no idea how things are ranked or why." He believes that in the "open society" he seeks on the Internet, it's important to have mechanisms in place that enable people to "look under the hood" at the tools they use.

It's not a complaint the average user of Google is likely to share. But given the extraordinary momentum of open source projects across the world of software and the internet, there is logic to his argument. He promises a public prototype of the search product "within the next few months," and "something really functional in a year or so." He already created one of the Internet's most useful and used sites in Wikipedia. At the least, it will be interesting to watch him and his merry band attempt to compete with Google (Charts), Microsoft (Charts), Ask, et al. I'll let you know if my daughter's teachers let her use it.