MediaGoblin: Saving the Internet Through Federation

By

A new project for sharing media files might be as important for its philosophy as its features.

In the last few years, services like Identi.ca and Diaspora* have popularized the idea of decentralized social media as an alternative to sites like Twitter and Facebook. Now, Chris Webber and Deb Nicholson are filling a gap in decentralized services with MediaGoblin, a project currently in early development for managing pictures, audio, and video. Like these other projects, MediaGoblin emphasizes its philosophy as much as the code itself.

“One of the things I hope,” Webber says, “is that MediaGoblin will not just solve the immediate problem of decentralizing media, but ends up to be a communication channel of why these issues are important.”

This mixed approach comes naturally to Webber and Nicholson, who are both long-time free software advocates. A veteran Python developer, until recently Webber was a software engineer at Creative Commons; he also created animations for the film Patent Absurdity. Similarly, Nicholson was a long-time employee of the Free Software Foundation (FSF). She remains one of the main organizers of the FSF’s Women Caucus and other parts of the FSF community.

Through their free software activism, Webber and Nicholson have been in touch for some years. When Webber first announced MediaGoblin, Nicholson immediately volunteered to serve as a communications consultant for the project. They quickly gathered a small, close-knit group – a fact that Nicholson jokes is due partly to the exchange of personal media for which MediaGoblin is designed.

“I never see any weirdness on the IRC channel,” she says. “It makes the community feel close-knit, because we’ve shared our cat and baby pictures already. I’ve seen a lot of pictures of kids in penguin undies.”

Features, Designs, and Future

Those interested in MediaGoblin can choose an instance on which to register. However, right now, relatively little is available to see. In its present form, MediaGoblin lets users create a profile of themselves; upload media one at a time; add tags, descriptions, and licenses for the contents of their portfolios; and not much else. More technically-minded users can follow the directions for setting up their own MediaGoblin servers. However, this basic functionality is being extended quickly. The API, as Webber put it, “just landed,” and the plugin structure is not much older (a sample plugin for ASCII art is the only one completed).

The current challenge, Webber says, “is to move beyond the bare minimum,” and the project is working hard to do just that. MediaGoblin is considering making the setup of its server easier – either through packaging, which Webber describes as “hard, because you have to configure the web server,” or possibly through virtual utilities. Also on the agenda are better administration tools, 3D model support, spam handling, and the conversion of some hard-coded features to plugins.

On the interface, Webber hopes to improve the uploader with drag and drop support, a progress indicator, and batch uploading. It is time, he says, to “to step back and look at the design cycle” after months of focusing primarily on adding the basic functionality.

The Philosophy of Federation

The development of such features should be largely routine. However, MediaGoblin’s largest challenge is to implement a more unusual feature – one called federation – that right now is part of almost nobody’s concept of the Internet. Yet, if projects like MediaGoblin have their way, it soon will be. Federation is an ambitious concept, aiming at nothing less than safeguarding against censorship and invasions of privacy created by centralized web services.

In fact, federation is such a novel concept that it requires educating users about why it is desirable, harnessing users’ concern about the control exercised by large sites like Flickr or Instagram to make them see the need for change.

As Webber and Nicholson explain in a promotional video for their upcoming crowdfunding campaign, the strength of the Internet has always been that it was composed of separate, but linked sites, many of which duplicated each other. This structure meant that if one site closed, either because of technical problems or censorship, the Internet could repair the damage by re-routing around it.

An example of a centralized, fragile Internet.

However, in the last five years, giant social media sites like Twitter or Google+ have resulted in many parts of the Internet becoming centralized – both controlled from a handful of sites and interacting with each other through the central sites, rather than directly with each other. This design is not only much more vulnerable to control, but also means that individual sites can disappear with many associated sites never becoming aware of the loss.

A particular problem is automated censorship – the removal of sites without human intervention. This problem exists because centralized sites are too large to monitor easily, yet their owners are concerned to act quickly to avoid any liability over possible violations of copyright or other laws. Instead of each case being examined separately, possible violations are removed first and examined later – if ever. As many users found out last year when they were barred from signing up for Google+ under pseudonyms, it can be difficult – and sometimes impossible – to find a human to correct the mistakes created by such automated mechanisms.

Consequently, Webber says, for regular posters of information, increasingly “It’s not a question of if my stuff is going to be taken down. It’s a question of when.”

One reason for Diaspora*’s initial success, he suggests, is that it was picked up by the media during a time when concerns about privacy on Facebook were widespread. “People have these freak out moments and want to pull out, then they realize they won’t be able to speak to their aunts, so they don’t bother. So, one of the things we’re trying to highlight is those moments of recognition that people have.” For a brief moment, Diaspora* seemed an alternative.

Part of MediaGoblin’s solution is its release under the GNU Affero General Public License, a version of the GPL specifically designed for transparency for cloud services. At least in theory, by making the code available, the Affero GPL will allow users to see just how a service might act when possible policy or copyright violations occur.

An example of a de-centralized, healthy Internet.

More importantly, MediaGoblin is dedicated to decentralized services. Their goal is not services offered by a single, centralized site, but, as Nicholson describes it, “a lot of independent sites that are working toward a similar goal, but they each have their own look and feel.” The arrangement is like that currently available for email: “Everyone has their own provider, whether it’s your personal domain email or whatever kind of email you choose, but then you are also able to talk to each other.”

Borrowing the term from Status.Net, MediaGoblin refers to this arrangement as “federation.” Being smaller and local, a federated site is less likely to resort to automated takedown or to enforce unacceptable policies. Should a federated site attempt such tactics, users have alternatives to which they can quickly switch that offer equivalent services.

In a sense, federation is an attempt to provide online services while restoring the old decentralized structure of the web. Yet, MediaGoblin’s intent is to go further than what has existed before, with search across all federated sites, including the ability to choose between multiple meanings of words in search filters.

Nicholson suggests that federation is a preferred alternative to that available from projects like Friendi.ca, which offers to return control to users by giving them a new site from which to control all the large centralized sites.

Although she finds Friendi.ca useful, Nicholson adds that, “I don’t think it’s the whole solution. It would be great for people to have lots of things to choose from, but there aren’t. There’s a couple. Most people think of one big site that they go to for all their hosting. What’s important now is that there should be multiple places where you can find media.”

In many ways, federation is central to the entire concept of MediaGoblin. For this reason, Nicholson and Webber emphasize it when discussing the project.

“Free software doesn’t always do as good a job as it could of speaking to non-free softwarish people. There’s an attitude that free software is just for free software-type people,” Webber says. “It’s important to send a message not only to free software people, but also to people who have never thought of these things before.”

Fundraising with the FSF

Building the software and explaining the need for it makes MediaGoblin one of the more far-reaching projects of the last couple of years. To meet the twin goals of coding and education, the project is about to launch a crowdfunding campaign, with a goal of allowing Webber to work full-time on the project over the next year.

Unusually, the campaign is being launched through the Free Software Foundation, rather than KickStarter or any of the other crowdfunding sites. “In some ways, it seemed a riskier decision,” Webber admits, but he feels that the advantages far outweigh the risks. For one think, the FSF was willing to customize the campaign page to MediaGoblin’s specifications. For another, with the FSF, there is “less risk of being lost in the crowd” than on KickStarter.

However, the most important reason for launching the campaign through the FSF is that “there’s a certain amount of trust when you go with the FSF.” Those who are likely to hear of the campaign are likely to have some grasp of the issues MediaGoblin tries to address and to need less explanation before they donate.

In many ways, MediaGoblin is still deliberately in stealth mode. “Before we start attracting people who aren’t into free software,” Nicholson says, “we plan on having something really big and glossy. [Right now,] we want people in free software and people who care about web freedom to know that we’re working on it. But we’re not ready for the harsh light of people asking, ‘Where are all the things I have on my Flickr account?’ We want to build a community with sustainable growth, not have everyone come in and see what we haven’t done, then go away and never return. We don’t want the fact that we’re unfinished as the distinguishing feature of MediaGoblin.”

Not that there seems much chance of that. With its combination of ambition and moral and communicational clarity, MediaGoblin is looking like a project that we’ll be hearing more about in the next few years.

Related content

comments powered by Disqus
Subscribe to our Linux Newsletters
Find Linux and Open Source Jobs
Subscribe to our ADMIN Newsletters

Support Our Work

Linux Magazine content is made possible with support from readers like you. Please consider contributing when you’ve found an article to be beneficial.

Learn More

News