Should the New Mantra Be: "Free as in Data"?

It's early at the Open Source Convention, held this week in San Jose — the keynotes were just this morning — but I'm already starting to see a few themes. One of them, mentioned both in the first keynote address this morning, is the notion that freedom is not necessarily synonymous with software freedom.

[ See also: Convincing the Boss to Accept FOSS ]

For example, said Tim O'Reilly, you can think of the "Internet OS" as a data operating system, based on the growing importance of some of its subsystems, such as location and identity — Who should own the databases? It's no longer an issue of having source code; it's the data you have and control that defines power.

"There's a natural tendency towards monopoly," said O'Reilly. And for those who care about freedom (of which software freedom is a subset), that raises the question: Should those databases be centralized or federated?

The meta-issues that question suggests, which touch on software as a service (SaaS), Internet architecture decisions, and "cloud" software licensing, were the topic of a panel discussion moderated by Bradley Kuhn from the Software Freedom Law Center, "With Software as a Service, Is Only the Network Luddite Free?"

Free software may not matter as much as it used to, said O'Reilly, and it may cause us to miss things that we ought to be paying attention to. "The architecture of systems matters more than the licenses," he pointed out. "The Internet isn't free because of a license but because it was designed for cooperation." It's not just the availability of the source code, because you don't have the machine and you don't have the data that's been collected. "How do we think of these services as fundamentally federated?" he asked. "We're going to live with the consequences [of the answer]."

Open source is based around the convenient fact that for 20 to 30 years, computing has happened on a user's local computer, said Benjamin Mako Hill, senior researcher at the MIT Sloan School of Management. One result is that open source philosophy and licenses are built around this highly individualistic use of software. But that's increasingly not the case in a world of network service, Hill said. "It's no longer as simple as a user's computing; it's a group's computing." And none of us is quite sure what is meant by "a group to control its computing." Some of these issues are addressed in the Franklin Street Statement on Freedom and Network Services (which I confess I haven't had time to read — hey, I'm listening to the speakers!).

One example of the new data challenges is how it is and should be licensed, so that people can collaborate but they don't give away more rights than they should. When you contribute to a social network, what do you grant to the public and what do you keep? If you leave the social network, what data can you take with you? That's one reason why O'Reilly believes the licensing model a developer adopts is among one of the most important decisions. "What is the license of my tweets? I really don't know," he said.

This isn't limited only to the information we consciously post on social networks, said O'Reilly. We also have to think about data that isn't created explicitly by humans. For example, your cell phone is automatically recording GPS locations; who has the copyright on that information? "As more companies wake up to the realization that data is power, they will not hesitate to make great use of it," O'Reilly added.

"We talk too much about free software and not so much about freedom," said Hill. "The value and principle at stake here is user autonomy."

Open standards are just part of the solution, said the panelists. Suggested Identi.ca's Evan Prodromou, an open source version of Facebook might be more integrated and you could control your own social graph on a server you control, without influencing the site's functionality. Because, he said, right now "If Romeo is on Facebook and Juliet is on facebook they will never be able to be friends."

"The low barrier to entry in the federated web is much lower when the development system is there," said Prodromou. "It's really important to develop and attract this kind of developer ecosystem," he explained, such as with APIs. Identi.ca's API supports the Twitter API, he said, "So we have great buy-in by third party developers."

But this is all about rethinking structured data and the way it's shared. "I'd love the FOSS community to reinvent the address book," said O'Reilly. "That's really my social network." In an ideal world, he said, the address book would understand e-mail, instant messaging, microblogging, and use algorithms and metrics to notice who's more important because he responds faster to them. "A social application like Facebook should access my machine using my settings of sharing to private/friends-only; the network would be with me," he explained.

There's plenty more discussion at the conference touching on the "what ought to be in the cloud, and who owns it?" Danny O’Brien, international outreach coordinator for the Electronic Frontier Foundation, gave a session about why innovation still comes from the edge of our networks during which he pointed out that the primary problem in giving data to the cloud is that most services are honeypots. "Even if they go to great efforts to protect your data, [these services] become the one stop shop for hackers, government, or their own corporate temptation."

They have no vested interest in protecting that data, said O'Brien. Yet we trust them to give us a feeling of privacy. "This is a strange and weird result," he said.

The conversation is obviously continuing... and it needs to.

Insider: How the basic tech behind the Internet works
Join the discussion
Be the first to comment on this article. Our Commenting Policies