Five years ago this month, Larry Sanger caused a stir when he launched a fork of Wikipedia, the site he helped create in 2001. Sanger believed Wikipedia had fatal weaknesses that could only be fixed by making a clean break from the past. He had devised a new editing process he hoped would allow his new site to eventually surpass the original.

It hasn't turned out that way. After a burst of initial enthusiasm, the site plateaued and then began to steadily decline. The site now has fewer than 100 active members, with only about a dozen of them making edits on a typical day.

To mark the site's fifth birthday, Ars talked to two people who have been deeply involved with the Citizendium project: Sanger himself and Hayford Peirce, a science fiction author who is currently a member of the site's editorial council. Both readily conceded that the site had fallen short of expectations. Read on for their candid assessments of what went wrong, and what we can learn from the experience.

"Little by little these people dropped out"

Wikipedia was Sanger's brainchild, but he left the project the year after it started. He disliked what he called the "ill-mannered geeks" who had come to dominate the site. And he came to believe that Wikipedia's completely open editing process gave insufficient deference to credentialed experts. He worried that if they were "forced to defend their edits on article discussion pages against attacks by nonexperts," they would burn out and stop contributing to the site altogether.

So Sanger announced a new project called Citizendium. It began as a fork of Wikipedia, but the site "unforked" in 2007, deleting imported Wikipedia articles that had not yet been modified on Citizendium.

Citizendium's launch prompted a great debate about the merits of Wikipedia's radically democratic editing process. In a widely-read essay, Clay Shirky attacked Sanger's fundamental premise that Wikipedia was hampered by its failure to properly defer to credentialed experts. Shirky argued that the key to Wikipedia's success was that it was focused on peer-review of edits themselves, rather than than the people making them. He predicted that Citizendium would get bogged down in fights about who qualified as an expert, which would waste time and generate ill will that would hamper the site's growth.

Shirky's predictions turned out to be prescient, at least if Hayford Peirce's telling of the Citizendium story is accurate. According to Peirce, the site attracted a number of prominent academics interested in contributing content. However, he said, these academics were "having long, ongoing fights with Larry about certain fields of expertise." Academics clashed with Sanger over esoteric topics such as animal taxonomy. "Do we call a lion a lion or a felix something or other. Professor types fight about it. Little by little these people dropped out," Peirce said. Today, only a handful of academics remain regular contributors.

Peirce also thinks that the site's processes were unduly complex, especially given the small size of the Citizendium community. "There were nominations for a charter-writing board and an election," he said. "There were some really strange people elected to it. They labored away for six months or so, and fought with each other, and a couple of them stopped working." Peirce says that by the time the charter was finished, there were only 40 to 50 people around to vote on it.

Peirce's story is backed up by Citizendium's official statistics, which show that the rate of article creation has declined precipitously over the last two years. At its peak in 2009, Citizendium boasted almost 30 new articles per day. Today, the rate has fallen to around 2 articles per day.

"We haven't given up yet"

Sanger is not as optimistic about Citizendium as he was on the site's one-year anniversary, when he was still predicting a "coming explosion of growth" in the site's contributors. He kept his promise to step down as site leader after two to three years in order to avoid becoming a "dictator for life." And he says he no longer contributes content to the site.

Still, Sanger is bullish on Citizendium's accomplishments. "We essentially proved that there is another way to use wikis," Sanger said. "Had Wikipedia started with a model somewhat closer to Citizendium, I think it would have been a much better project, both in terms of management and quality of content."

What went wrong? Sanger argues that the primary problem was Wikipedia's overwhelming first-mover advantage. "As long as Wikipedia is a top ten site, it's going to be difficult for a competitor to get any traction," Sanger told Ars. "A lot of people are going to try to draw the conclusion that there's something about the model which meant that it couldn't take off. I really don't think that's the case. The model works very well in many ways."

Today, the prospect of Citizendium overtaking Wikipedia seems pretty remote. Was Clay Shirky right that Citizendium was undone by its unwieldy editing model, or was it doomed from the start by Wikipedia's first-mover advantage? I've long been in the Shirky camp, and Peirce's comments lend some credence to Shirky's predictions. At the same time, it's hard to say Sanger got a fair test of his model. By 2006 Wikipedia had so much momentum that it would have been hard for any alternative, no matter how well-designed, to gain traction.

Either way, Citizendium's future looks grim. "I personally think the site is going to fail unless someone like Larry comes up with a sugar daddy who will be able to pay 300, 400, or 500 per month just to keep the servers going," Peirce told Ars.

"We haven't given up yet," counters Sanger. "We don't intend to any time in the foreseeable future. As long as there are people interested in building Citizendium, I'm going to make sure, one way or another, that the servers stay on."