Distributed Social Networks and Public History

· Posted by Joshua in Online History

Diaspora* cofounders Daniel Grippi and Ilya Zhitomirskiy.
Daniel Grippi and Ilya Zhitomirskiy, cofounders of the distributed social network Diaspora*, present at the Cultura Libre Conference in 2011. Flickr Photo © 2011 Jalisco Campus Party, CC BY 2.0 License.

Distributed social networks are not well-known, but their tumultuous history offers a range of opportunities and challenges to public historians on the web. Distributed networks, such as GNU Social, Diaspora*, or Friendica, exist as an alternatives to more popular sites like Facebook, Twitter, or Tumblr, which are centralized networks. Each centralized network exists on a single website, owned and controlled by a single business with whom all users of the network must register accounts. Distributed social networks, by contrast, run on free community-built software that can power any number of different access points. This allows users to join the network from their choice of sign-on providers, or even to access the network by creating a server of their own, without relying on a single centralized authority.

Public historians can learn from both the aspirations and failures of distributed social networks. For historians who aim to collect digital history, these fragmented, privacy-conscious networks pose a dilemma about how to preserve online communications while respecting user privacy. For museums and cultural institutions who aim to build sociability into online exhibits and collections, distributed networks offer an ethical model for open development, interoperability, and privacy protection — as well as case studies in the difficulty of attracting engaged users.

History of Distributed Networks: GNU Social, Diaspora*, Friendica

Distributed social networks began to appear around 2008-2010 in response to fears that centralized networks threatened their users’ privacy and autonomy. The companies that own centralized networks like Facebook or Twitter dictate lengthy terms of service that can change at any time, freely utilize users’ data to sell advertising, and create a single point of failure giving users little recourse if the company suspends an account, merges into another company, or goes out of business. Distributed networks promise to give users the ability to access a network without being locked into a single company’s product — just as a telephone user with AT&T can call someone with Verizon, or a Yahoo! Mail user can send a letter to a friend at Gmail. This empowers distributed network users to choose the provider whose terms they like best, to switch providers while still being able to socialize with the same contacts, and even to set up their own service for fine-grained control over their personal data.

One of the first distributed social networks was GNU Social (first called Laconica and then StatusNet), created by Evan Prodromou in 2008. Prodromou explained at FOSDEM in 2010 that while social networks had proliferated over the 2000s:

Status updates from different parts of the web don’t come together and this mesh of sociality is not completely connected — and these systems are by and large closed. … We need social software that we can own, that we can install and manage ourselves. We need social software that meshes into the existing social web.

GNU Social offers a Twitter-like service for sharing status updates across multiple websites. It still powers sites such as quitter.se and gnusocial.no, although in 2012 Prodromou turned to a new project called pump.io, which introduced additional features for image sharing, social gaming, and long-format text. Pump.io remains under development, but it already powers identi.ca, microca.st, and other sites. Because it is free software, users are also encouraged to install and manage their own pump.io server to connect with the network — an emphasis that may intimidate those without a tech background, slowing the network’s growth.

A more widely known distributed social network is Diaspora*, which was initiated in 2010 by New York University Courant Institute students Ilya Zhitomirskiy, Dan Grippi, Max Salzberg, and Raphael Sofaer. They created Diaspora* out of a concern that centralized networks, particularly Facebook, were collecting too much personal information and spying on users. As Salzberg told the New York Times:

When you give up that data, you’re giving it up forever … The value they give us is negligible in the scale of what they are doing, and what we are giving up is all of our privacy.

Diaspora* empowered users to access the social network from their choice of “pods” owned by different hosts, or to install their own “pod,” preventing any single entity from controlling the network. Diaspora* was also among the first social networks to allow users to create separate “aspects” (or groups of friends, family, or coworkers), in order to choose expressly which contacts could access different profile information. These features drew significant press attention and allowed the Diaspora* project to quickly gain over $200,000 in donations via the crowdfunding service Kickstarter. Development of the project was slower than anticipated, however, and trouble compounded in 2011 when PayPal froze the projects’ funds without explanation and co-founder Ilya Zhitomirskiy tragically took his own life at age 22. The remaining founders later handed development of Diaspora* to the open-source community, and the platform today powers over 300 pods with more than 600,000 users. Even so, Diaspora* has never lived up to the initial high expectations that it could be a “Facebook killer.”

Distributed social networks have struggled to attract users in part because of the “network effect” — the more users a network has, the more useful it becomes, since people can already find most of their friends and associates there. This makes it hard for new networks with few users to ever catch up to a large established networks like Facebook.

Former Netscape programmer Mike Macgirvin aimed to overcome the network effect in 2010 with a federated social network called Friendica (also briefly known as Mistpark and Friendika). Users who installed or signed up at a Friendica server could communicate not only with other Friendica servers, but also GNU Social and Diaspora*, as well as Facebook and Twitter (provided users still had accounts with the latter proprietary services). In theory, this would make it easy for users to start out at the distributed network without leaving behind contacts at centralized services. Friendica attracted thousands of users and remains active, but Facebook and Twitter have frequently broken compatibility by changing their APIs or terms of use, requiring Friendica’s developers to constantly play catch up to maintain compatibility.

Ultimately, GNU Social, Diaspora*, and Friendica are still mostly the province of technology and open source enthusiasts. Their user base reflects the technological barriers to getting started with these platforms, the relative lack of marketing apart from word-of-mouth discussions among developers, and perhaps the tech community’s heightened perception of the dangers posed by centralized networks.

Lessons for Public Historians

Distributed social networks have implications for public historians who aim to collect digital records as well as those who seek to use the web to encourage social engagement with history.

As the very term public history suggests, public historians are usually more focused on publicity than privacy. They aim to share history with the widest audience, or to collect the past from the widest public. The proliferation of distributed social networks, however, has been predicated on giving web users more individual control over their privacy and their data. Historians must be respectful of personal preferences about privacy, while also reminding the public that privacy controls could leave valuable historical information inaccessible in the future.

Although it represents a small segment of the population, much of the speech posted to distributed networks is historically significant for the way it contests privatized, corporate control of the world wide web. In some sense, distributed networks constitute a protest movement against the commercialization of online sociability since the 2000s. Distributed social networks have also become platforms for politically-charged speech that risks censorship on centralized services. We may view some of this material, such as ISIS activity on Diaspora*, as reprehensible, but that makes it no less interesting for future historical study. Distributed networks may or may not make such content harder for historians to access — control over content has simply shifted from the network providers to network participants. Digital archivists must navigate between the more open protocols of distributed networks and the privacy controls enabled by their users as they document this history.

Public historians who aim to integrate social features into new web exhibits and collections can also learn from distributed social networks. The way that these networks give participants ownership over their contributions and allow interoperability between many access points provides an ethical model that historians may want to emulate. In particular, public historians should aspire to promote connections between different institutions and their patrons, rather than allowing exclusive walled gardens to proliferate at every museum, archive, or historical society. Since distributed social networks typically already run on free, open source software, historical institutions could even modify this software to power their own interconnected online communities. Whatever the technological implementation, such interconnection would require concerted collaboration between institutions, as well as willingness to share authority outside institutional walls. The benefit, however, would be a larger, more democratic network for everyone, where cultural heritage is freed from hard-to-find individual websites and institutions and distributed among members of the history community wherever they might be.