When Jean-Baptiste Kempf joined École Centrale Paris as a student in 2003, he was tasked with helping run the university’s computer network. It included an unusual project: student-run open-source software that had been running on a couple of university servers for seven years. To students, the project was known as “Network 2000.” To the rest of the world, it was VLC media player.
Kempf—now the president of VLC’s parent organization, the nonprofit VideoLAN—is the person who helped guide VLC’s journey from student project to ubiquitous software. (VideoLAN Client, the original name for the project, is where VLC gets its name.) On the surface, he’s laid-back, casual, and frank, though that belies a steely determination. As the person overseeing the project and its team, he sets the tone for VLC as a whole.
VLC is a juggernaut among media players. Since February 2005, it’s been downloaded 3 billion times, according to VideoLAN. Even before the nonprofit began tracking downloads, it was clear that VLC was a runaway success. In an era before Netflix and YouTube, users had few choices when it came to media software, and none of them were terribly good. There was Windows Media Player, a janky and underpowered program sufficient for entry-level users, as well as monstrosities like RealPlayer, which locked users into odd codecs and file formats. Though first developed in 1996, VLC was a breath of fresh air when it was released in 2001 under a GNU General Public license: It was customizable and high-powered, and, above all, it worked. That release extended VLC’s reach beyond École Centrale as the software was adopted by the wider public.
In an era before Netflix and YouTube, users had few choices when it came to media software, and none of them were terribly good.
But by the late aughts, VLC was facing a terminal diagnosis. The French university had first developed it as a way for computer science students to practice their coding in a safe environment. Initially, the school threw a lot of resources behind it, folding it into a formal teaching program. But VLC player became a victim of its own success: It outgrew its server, became too complicated for a succession of final-year students to maintain, and was soon outmoded by a curriculum change.
By 2007, “nobody was really taking any care of it,” says Rémi Denis-Courmont, a lead developer on VLC. He had started working on the project in November 2003, as a student. “It was a bit of a hot potato,” he continues. “It was more than they wanted to handle or could handle.”
“The project was dying at the university,” Kempf says. That’s when he made a decision: In order to survive, VLC had to cut ties with École Centrale.
In the years since then, VLC has undergone significant changes. Having survived for more than two decades, with many of its current team in place for 15 years, the group has evolved from a small student project into a nonprofit and a consultancy, all while continuing to offer the free, open-source software that remains their raison d’être. The project has persisted in large part thanks to the organizers’ ability to adapt to the times and to make hard choices, like Kempf’s decision to lead VLC’s departure from the university rather than watch the project peter out.
VLC had always been run by a small core team: Two to five people constantly worked on the project, and about a dozen others chipped in semi-regularly. “When I started to get very involved in the project [in 2007], there was only one person really involved in the project, or one and a half,” says Kempf. (That one was Denis-Courmont.) “We were two and a half with me.”
That core of key contributors made Kempf’s plan to leave easier. In 2007, he called everyone currently and formerly active on the project. He then laid out the situation: The university could no longer support a project of this size, and the stream of student contributors was drying up. Moreover, the success of VLC player had made it impossible—even under ideal conditions—for the school to keep the project up alone.
It wasn’t a difficult decision at all, remembers Denis-Courmont. “We didn’t really have a choice,” he says. “The university didn’t want to host it anymore.”
Instead, the challenge was deciding how VLC would move forward— and that was more difficult.
Discussions were held the week before Christmas 2008 at the project’s first-ever Dev Days conference. “We said, ‘Let’s do a nonprofit,’” remembers Kempf. He later filed the necessary paperwork. There was little change to the way the tight-knit team worked, and the public using the program saw no difference whatsoever.
The creation of the nonprofit was the second major move for VLC in its short history. In 2001, after a long fight between the university and the custodians of the project, École Centrale Paris relicensed VLC under the GPLv2 license. This allowed people outside the university to contribute to the now open-source project for the first time—which resulted in speedy ports to Windows and MacOS.
Opening the doors to VLC meant broadening its contributor base and diversifying its core team. Felix Paul Kühne, currently one of the longest-tenured contributors to the project, started working on VLC in early 2003, when he was a 16-year-old high school student in Germany. At the time, VLC was one of the first media players available for Mac, and was still hosted on university servers.
“They had a website where you could download VLC, and they asked for contributions,” recalls Kühne. But at the time he couldn’t code. “One of my first big contributions was a German translation.”
He emailed the core VLC team a word-for-word translation of every dialog box and all the menu text in the program. It was a good in: “Over time, I had university students who were five years older than me and way more skilled in coding and writing software talking to me,” he says. “We got in touch on chat platforms and discussed features to add, and so I learned how to code.”
Kühne, now a doctor of emergency medicine in Bremen, remains a lead developer on the project. He devotes his spare time to corralling those who contribute to the project, while adding his own commits to VLC’s gigantic codebase. That combination of committed amateurs and full-time professional staff members shows the curious conflict at the core of VLC: Even though it’s a massive program used by millions worldwide, those tasked with its upkeep are keen to keep everything small-scale.
“The core team is maybe 10 people, but every year we have 100 or 150 people contributing,” explains Kempf. “That means that, statistically, people who send the changes are never going to maintain the features they send.” Major decisions are kept within a tight group of voters, who are part of the nonprofit associated with VLC.
Another challenge is the way in which VLC accepts contributions. While anyone can submit new ideas, whether they’re accepted or not largely depends on the quality of the code, rather than the utility of the feature. “We have a very difficult process,” Kempf admits. “People don’t understand why we do that. But the reason is we need good code, because I’m the one that’s going to maintain it. If I don’t understand it now, this is going to be a problem.” While Kempf acknowledges that the high barrier to entry discourages some from contributing, he’s unapologetic about being hard on code quality.
This makes VLC a target for naysayers, Kempf reckons. “People say, ‘Why don’t you have this obvious feature that everyone should have, because it’s obvious that every other player has it.’ We’re like, ‘Well, no one managed to do it in a nice way.’”
Such an attitude has meant that some odd features—such as the puzzle module, which turns any video being shown through the app into a playable jigsaw game—have persisted since version 0.9.0 of VLC player, which was released more than a decade ago.
“There are some [silly] features in VLC—some filters that everyone finds weird and asks [us], ‘Why do you have that?’ The answer is someone sent a good patch, so there was no reason to refuse it,” says Kempf.
Tech writ large relies heavily on open-source software—it’s behind the majority of web browsers and media players, as well as a good chunk of the code that makes our operating systems work. However, as Kempf says, “it’s difficult to tell new and young people that you should spend your days working for free or not much.”
Yet the team still manages to attract new blood, in part because of VLC’s prestige. “Saying you’ve worked on VLC gives you some good pedigree,” says Kempf. Altogether, more than 1,000 people have contributed to the project over its history.
Still, in the early 2010s, VLC wanted to make changes— like their expansion away from desktop—that required more formalized effort.
“We wanted to be on smartphones and smart TVs and everything related to those technologies,” says Kempf. “We couldn’t do that just in our free time. Putting stuff on the [app] stores is very time-consuming. It requires a lot of hardware, and a lot of that hardware is very boring.”
Kühne adds, “We [also] had a noticeable amount of companies asking for commercial support or specific features, and [when we were first involved in the project] all we could tell them was, ‘Yeah, sorry, we’re students, we can’t help you.’”
In 2012, “the idea was born to make a consulting company” connected with VLC, Kühne says. A for-profit arm of the organization, called Video Labs, could also help with VLC’s outstanding projects. (And, Kempf says, “startups are hot now, not free software.”)
“When we started VLC, we didn’t imagine the iPhone or iPad would happen.”
The decision split the VLC contributors, but Kempf managed to convince them when he explained how the interaction would work. “The company is doing all the boring tasks,” he says with a laugh. “There are some important refactorings that are crucial to do and no one wants to do them, but when you start paying people, you can say, ‘You know what, do it.’”
Since 2012, the company has employed a small cadre of paid staffers who do the jobs that committed amateurs can’t or won’t do: Employees completed most of the work needed to get VLC onto smart TVs and cellphones, which required recoding much of the app from the bottom up. Separate from the nonprofit, which is staffed only by volunteers, the company’s workforce has tackled the majority of app development.
“It works. VLC is still available for free, with all the core liberties, and [remains] open source. The development process is completely in the open,” Kempf says. The commercial arm goes “unnoticed for most of the users.”
The consulting company has also helped VLC adapt to mobile. “When we started VLC, we didn’t imagine the iPhone or iPad would happen,” says Kühne. “Right now, we have almost the same number of users on Android or iOS as we have on the desktop.”
Read more in issue 9
This issue explores the inner workings and wide-reaching impact of open-source software—and the communities that build it.
But managing a disparate group of volunteers and workers across a nonprofit and a for-profit company is difficult. Communications are managed through an IRC channel, which has grown increasingly quiet over the last few years. Most of the momentum is maintained through in-person meetings.
Denis-Courmont is worried about the future. “We’re just not attracting the traditional open-source contributors—[whether] students or employees from large companies,” he says. “At the moment, it’s basically me and Felix [Kühne] and maybe one other person doing it in their free time,” alongside the paid company employees.
To try and counteract the downturn, the team members behind VLC plan to change their main forms of communication from IRC chat and a mailing list “to a more modern approach where you can upload your stuff to a website and do annotations not in email but on a proper platform, like in Github,” says Kühne.
They’ve also made VLC far more accessible and welcoming for people who contribute. The team has built out a suite of documentation, and made it easier to both build Docker containers and use tools. Inspired by the concept of bug bounties for virus hunters, the nonprofit has also, since 2011, directed some money toward rewards for people who contribute high-demand features.
The group meets twice annually for an all-hands meeting to discuss the broader movement and development of VLC media player: once at an open-source software conference in Brussels, during the first week of February, and, since 2012, at VideoLAN Dev Days, where roughly a hundred attendees come together to discuss the future of the project. These gatherings allow the disparate, global team behind the player to build a community akin to the one that evolved more than a decade ago, over long, hard days spent coding on campus.
What does the future hold for VLC media player, the odd university project that snowballed into a desktop stalwart and a curious combination of open source, nonprofit, and consulting company? VLC isn’t YouTube, says Denis-Courmont, but “we still have millions of users. It’s not going to vanish.”
“I have no idea what it’s going to be like in five or 10 years,” Kempf adds. “But I think we’re developing good technologies that can be useful.” And more importantly, by striking a happy medium between a slender startup and a sprawling, open-source software project, VLC’s leaders ensure that its contributors feel like a valued part of the team and that they’re pulling in the same direction—to produce what is arguably the world’s most beloved media player. In this way, they’ve managed to both resuscitate a small student project and grow it into the giant platform it is today. Together, they’re the team that powers the player.