X

Wikipedia's open-source label conundrum

Despite a media habit of referring to Wikipedia as an open-source project, it looks only a little like such software efforts.

Daniel Terdiman Former Senior Writer / News
Daniel Terdiman is a senior writer at CNET News covering Twitter, Net culture, and everything in between.
Daniel Terdiman
5 min read
The online encyclopedia Wikipedia is often referred to as an "open-source" project because it is written, edited and policed by a global group of volunteers.

However, the open-source label doesn't really fit Wikipedia. "Free-for-all," in fact, may be a better match.

"Open source," at least the way it's been used in tech circles over the years, usually connotes successful, volunteer projects like the Linux operating system, which has strict controls and is monitored by a handful of people who make the call on what is handed over to the public.

That hardly describes Wikipedia, whose own founder doesn't even like to call it "open source." With Wikipedia entries, no one calls the shots, and anyone, even anonymous users, can make changes to the publicly used product.

Simply put, the kind of confidence attached to Linux shouldn't be attached to Wikipedia, which is more a grand and very subjective experiment in collective writing than a rigid engineering project.

Two scandals have made that point clear. Last week, a former journalist named John Seigenthaler published an op-ed in USA Today blasting the service for an anonymously written article that had been on the site for four months linking him to the assassinations of both John F. Kennedy and Robert Kennedy. Then, former MTV host and podcasting pioneer Adam Curry was accused of editing out references to fellow contributors to podcasting technology in the Wikipedia article on the subject.

In response, The New York Times, for one, has banned reporters from using the Wikipedia as a research resource, according to a posting Wednesday at Poynter Online.

The editors at the Gray Lady who made that decision have a good point. Unlike Linux, Wikipedia has no central editor like programmer extraordinaire Linus Torvalds and his small cadre of managers, who make the final decision on what goes in the software. Wikipedia is always being modified, so there's never really a "production version" that's been tested and deemed reliable. And because Wikipedia makes it easy for contributions to be made anonymously, it lacks the accountability that forces open-source software developers to offer up their best work.

Credit and accountability
Most important, programmers relish taking credit for their contributions. That gives them credibility among other coders, makes them accountable when they produce something that doesn't work, and maybe even helps them land a good job.

"It takes two years to get 'street cred' in Linux software development," said Matt Asay, the founder of the Open Source Business Conference, a series of business conferences on open-source technology. "The time frame might be different with different projects, but the system is basically the same: There are only small groups of submitters. And it all has to be filtered through captains or those who have final access to the code."

By comparison, Wikipedia had 45,531 registered members at the end of October, 1,854 of whom made more than 100 edits during that month alone.

In fairness, though he rightly believes Wikipedia has some of the attributes of an open-source software project, Wikipedia founder Jimmy Wales doesn't like it when he sees someone describe it that way. "I think I've lost this battle against the term," Wales said.

There are other big differences between Wikipedia and open-source software.

Almost anyone can create a Wikipedia article that's made public, even though on Monday the service implemented a new rule requiring article writers to register with the site. Registering takes just a few seconds, and all Wikipedia asks for is a user name and password.

In fact, many posters do say who they are. But a few do not. Unregistered users can still edit any existing entry. That was the case with the anonymous poster who created the bogus Seigenthaler entry and Curry's embarrassing foray into editing. He was discovered when the edits were traced to an IP address for his computer.

If edits are incorrect, they won't get changed until someone else makes the changes. So in effect, there is no final version of a Wikipedia entry. They're more like living documents, always subject to change. Does that mean Wikipedia is always wrong? Not at all. But it does mean readers should be cautious.

Wales himself can make things confusing. As he tries to separate Wikipedia from the term "open source," he's quick to say Wikipedia--in addition to the volunteer force--shares several attributes of open-source software projects: The ability to copy, modify and redistribute. Also, like many open-source software projects, people are free to create derivative works from Wikipedia.

"I would say we are absolutely, definitely, open source, by what people mean (when they use the term)," said Wales. "But I don't use that term. Instead, I would say that we are part of the free culture movement" that also includes Creative Commons, the nonprofit organization that has created a licensing framework that allows for broad noncommercial use of copyrighted material.

But accountability is still missing, and that's a big difference from what goes on with software. The open-source software community itself, which is generally supportive of Wikipedia, is not so quick to agree that the terminology should apply to Wales' project.

"With software, you get the benefit of there being some objective measure of a change," said Brian Behlendorf, co-founder of the Apache Software Project. "It's not a question of aesthetics. Does this fix the bug people have noticed? Does it speed up performance?"

Other open-source experts agree that Wikipedia's lack of centralized control sets it apart from the software model.

"With open-source software, there tends to be an organization or person who's a leader of it and who's responsible for things going into the main line," said Dave Rosenberg, a principal analyst at the Open Source Development Lab, which promotes the adoption of Linux in the enterprise. "For instance, if you write a patch, Linus (Torvalds) approves it before it goes into Linux. It's the same with Apache."

Wales acknowledges the difference.

"Over time, as code gets to a certain level of maturity," he said, "it moves to a stable branch. And we're not at the point where we're ready to have a stable branch."

Still, Asay thinks anonymity is an inherent Wikipedia weakness.

"You can't do that in open source," Asay said. "You have to be a known entity to contribute code. You should have to be a known entity to contribute to Wikipedia."