Last week, Facebook announced that it had amassed 500 million users, a formidable portion of the global Internet audience. But even as Mark Zuckerberg and company celebrates, others are busy trying to uproot Facebook's popularity by establishing a set of open standards to share Facebook-like features across the Internet.
Just like open standards for e-mail and the Web broke users free from proprietary closed networks of the early 1990s, so too could a new set of standards allow people to share their thoughts, photos and comments across the Internet, regardless of what social networking services they use, argued Evan Prodromou, head of open source microblogging software provider StatusNet, during the O'Reilly Open Source Convention (OSCON), held in Portland, Oregon last week.
Open-source social, or "open social," networking services are not new. StatusNet has been running an open source implementation of its Twitter-like microblogging service for several years, called Indenti.ca. But no open-source service has gained Facebook- or Twitter-proportioned success.
Now, the developers behind such services are changing their pitch: Instead of stressing the open-source nature of their services and software, they are emphasizing how the interoperability of such offerings could free users -- and their data -- from the locks of any one social-networking service.
Prior to OSCON, a number of social-networking software developers gathered for an informal summit to discuss interoperability. They developed a simple test case to show how federation of social-networking services could share data.
In their example, a person uploads a photo of another person on some photo-sharing service, tagging the photo with the subject's name. The subject of that photo should automatically see the photo on his or her own preferred photo-sharing service. A friend of these two individuals who uses yet another service could see the photo and add a comment, and the message can then be relayed to the two other services.
"A federated social network would be a network of networks, using open protocols and a uniform name space that would allow anyone to participate," Prodromou said.
Such interoperability should be an inevitability, given the history of the Internet, Prodromou argued. Once some company-specific commercial technology gets really popular, it tends to be replaced by a set of open standards that multiple service providers use to offer generic versions of that feature.
E-mail is one example of this. "E-mail in 1992, 1993 was characterized by separation. We had large consumer services like CompuServe and Prodigy, with millions of users," Prodromou said. "It was used as a retention mechanism. You had to be on AOL [America Online] to e-mail someone on AOL." Governments and universities and single operator bulletin board systems (BBSes) also offered e-mail, though it was difficult relay messages across different systems.
But within two years, almost all these parties had switched to using the Internet for e-mail, deploying open standards like the Simple Mail Transfer Protocol, and open source software like the Sendmail SMTP server, Prodromou said.
"Federation maps very well to the way the Internet works," he said, also referring to document sharing through the Web and global services such as the domain name service.
In each of these cases, there was a great demand for functionality, and, as a result, multiple vendors used some open standard to offer that feature. And the commercial walled gardens of the dominant players gave way to open Internetwide availability of such services.
Take, for example, America Online (AOL). Those who were around in the mid-1990s remember when AOL exploded in popularity, thanks in part to the way it offered easy (though limited) access to the Internet features like e-mail, chatting and Web surfing.
"AOL would be the first, formative [online] experience for tens of millions of users, noted David Cassel, who ran the AOL Watch Web site in the 1990s. Like Facebook today, it was criticized for its policies on privacy, security and willingness to sell user data, as well as its use of its own proprietary tools.
Over time however, AOL's influence seemed to wane as more people signed up for broadband connections, which offered more direct access to e-mail and Web surfing. While still a prominent media company today, AOL is hardly the dominant online portal it once was. Could the same thing happen to Facebook?
Prodromou recounted a few of the recently developed standards that could bring about, in his words, a "federated social Web."
OpenID and OAuth could be used for authentication. Web identities could be established through Webfinger and Portable Contacts. Protocols such as Activity Streams and Google's Pubsubhubbub can be used to publish events and notifications, so that when a notice is published by one service, other services can be notified. A protocol, called Salmon, can be used for submitting comments across multiple services.
From these protocols, entire social networking stacks could be built, Prodromou argued. One existing stack, called OStatus,combines Pubsubhubbub, ActivityStreams, Salmon, Portable Contacts, and Webfinger into a single integrated package, which could be used to build photo sharing sites, travel sites, or any other sort of site with social participation.
Elements are still missing, Prodromou acknowledged. Privacy is the biggest factor: All of these protocols were based on the idea of sharing. In addition, a way of indicating who can see what is still needed. Application programming interfaces are needed as well, so material could be shared across multiple platforms, such as mobile phones.
Nonetheless, a number of open-source social networking services have been created, or in the process of being created, with these protocols. Perhaps, the highest profile project has been Diaspora, which has been featured in the New York Times. The developers behind Diaspora, which uses the OStatus stack, plan to unveil the service in September.
Other open social Web applications, all in various stages of completion, include Diso, the Open Source Social Networking Engine, GNU Social, BuddyPress, Vodaphone's OneSocialWeb, Appleseed and Crabgrass.
Moreover, groundwork is being laid to formally ratify these protocols. The World Wide Consortium (W3C) is looking at ways to develop standard protocols for social Web activities, and is considering ratifying existing standards like OAuth. A W3C Social Web working group plans to issue a report next month on the feasibility on establishing standards.
Despite this plethora of development activities, the jury is out on whether the world would be ready for a federation of interoperable social networks.
"The demand isn't really that clear. The problem with social networks is that there are only one or two [that are popular now], not thousands. So I don't think the desire is there to intercommunicate," said Chris DiBona, the open source manager for Google. He also said that the privacy issues would also be a concern.
DiBona noted that standards would be come more popular if more popular social networking sites crop up, or the internal ones run by enterprises grow more popular. "I think it is nice that people are trying, but I'm not sure the analogy works," he said, referring to comparisons of social Web protocols with those for e-mail and Web standards.
"It is hard to dislodge incumbents and big walled garden networks, but this also means there is lots of opportunity for small players to participate," Prodromou said. "Only once the federated system network becomes ubiquitous, do the big networks move in."