Decentralizing: ending walled gardens
Beyond the walled garden
If you’re a disgruntled user of social media, or an entrepreneur or investor seeing opportunity in the wave of disenchantment, you’re wondering if alternative social media are important, or is “just” creating more walled gardens a bad idea? Instead, do decentralized intermediaries, such as Twitter’s BlueSky, ActivityPub and Digi.me, show the way forward? Historic analogies suggest that they may be crucial, not just for salving Facebook’s specific wounds, but for enabling a future, well-interconnected system of autonomous social media.
— -
Some of my friends are closing their Facebook accounts; others are sternly defending their decisions to remain on Facebook, despite the grave concerns they acknowledge about the firm’s business and ethical lapses. Decentralized technologies are — increasingly — offered as the basis of new social media.
Who cares? As I update this article, on 14th January, 2020, FB stock is just about at an all-time high ($219 / share, $624B market cap), and with at least one set of analysts writing that it remains undervalued ( despite a nose-bleed price / earnings ratio of 35).
There’s a lesson from history, I think, that suggests that we’ve already passed the peak of the scope and impact of monolithic, monopolistic social media entities, of mutually incompatible walled gardens. We’ve seen this movie before. Several times.
A short history of walled gardens
Long ago, before any of us were born and in a distant time without computers, there were overlapping telephone networks. Incompatible networks. Meaning: if you wanted to place a call (quaint old language) to the Smith and Jones law firm, you first needed to know what network they were on — and you needed to be on the same network.
- In 1899, The Atlanta Telephone Company and the Southern Bell Telephone and Telegraph Company both offered service in Atlanta. “The two competing companies did not provide interconnecting service. You could only call other customers service by the same company. This forced most retail businesses to have dual service, listing both telephone numbers on their advertising. Some businesses managed to get the same telephone number on both systems.”
American Telegraph and Telephone Co., now AT&T, known for decades for its sharp-elbowed business practices, would refuse to interconnect between its own customers and those on other networks.
And again: in the 1980s, as mobile telephony emerged, it seemed that each major country would have its own homebrew standards. At the time this was annoying more than awful. It was of little concern to people who did not need to travel across national borders. If your business was entirely within Japan, or within the USA (a latecomer to mobile telephony), that wasn’t a problem. But in Europe, where proud nations with their own terrific universities created national standards, the consequences were that these conjured standards that wouldn’t work if people travelled a short distance within Europe, and that was a debacle in the making. Europe led, then, in the creation of GSM, (Groupe Spatiale Mobile), the first standards that took the priority of universal access to be higher than domestic pride.
Yet again — I check my own smartphone and, here we are in 2020 and I see the following message apps: SMS, Google Duo and Hangouts, Signal, Facebook Messenger and Whatsapp, Skype. Really. I don’t use Kik, WeChat, Snap, TikTok, Telegram or any of a long list of others. None of us can send a message from any of these to a user on any other. Your mix will differ, but it’s all just crazy. In Korea, Japan and Taiwan: you have to have Line to be taken seriously for messaging (partly because of the sophisticated use of Emojis). There are others prevalent in the Middle East (Botim, for example, and ToTok — which apparently doubles as government spyware), and so on.
As concerns grow about Facebook’s clout, would-be competitors include Jimmy Wales’ WT:Social, Mastodon, Planetary, MeWe, Matrix. Surely, adding to the list of would-be competitors to Facebook doesn’t on its own solve the problems anymore than building a third or fourth unconnected network in Atlanta in the late 1890s would have challenged AT&T.
Adding walled gardens doesn’t solve the problem of walled gardens.
The problem grows
If you live, as I do, in the USA (or allied countries) you may see Facebook’s global position as unchallenged. It isn’t. For reasons that aren’t all pretty, China bars all Facebook tools, and has its own native apps (Chinese users typically have more full integrated application platforms rather than fragmented apps — WeChat has chat and Facebook-like social media, and over 1.1 billion users). And similarly, Russia, for often similarly non-pretty reasons, favors VKontakte.ru with several hundred million users.
So, it’s 2020 and you live in the West but you’ve friends, from college, work and family, say, in the USA and China and Russia, and you have a new baby, or a great vacation with fantastic pictures, or want to share sad news about your cousin’s untimely death. So much for modern technology: you can’t do that with a single posting, or via a single means. And your friends and relatives around the world can’t join in the collective congratulations, joy, or mourning.
We’ve created the equivalent of the 1890s telephone network.
Why have China and Russia built their own walled gardens? Why are Facebook users more reticent in their sharing?
One major concern is about the data about individuals, the users, and how these data are used. The Chinese government (and Communist Party) demand control over the data about Chinese residents and citizens and are not interested in sharing that with anyone else. Ditto the Russian government. And Facebook’s seeming nonchalance about data scraping for nefarious purposes gives even its most ardent users reasons to pause.
That’s not the only reason. To be clear, China isn’t only concerned about outsiders having access to data about its people. It also is insistent on blocking discussion of many subjects it considers off limits: the three “T”s (Tianmen Square, Taiwan and Tibet), human rights in general, the uprising in Hong Kong, repression and imprisonment of Uighurs, any criticism of the Party and President-for-life Xi, and any reference to Winnie-the-Pooh. Online entities generally blocked from access within China include the New York Times and Amnesty International, so there’s not going to be any social media feed in China from these or similar entities.
But let’s set aside, for the purposes of this note, the content-blocking. Let’s consider only the incompatibility issue.
Interconnecting: paths between gardens
I believe that the best answer for users — and potentially for service providers — isn’t breaking up Facebook’s near-monopoly. It’s instead creating an interconnect layer for data sharing.
Suppose you and I are happily connected on one social media network, where the word ‘happily’ explicitly includes that we are content with its business model and practices. Okay — our cat videos are shared merrily as they are today.
But what of our need to share these with others, worldwide? It is here that a utility-form interconnect layer can exist, with APIs such that a worldwide publishing of your cat video could be viewable to all. But the rich longitudinal data on how individuals — including those viewing it — behave (other than, perhaps, applauding the video) is not shared at all beyond the confines of each viewer’s service.
The high-level concept is that multiple users can be either connected directly via their own chosen social network, which has its own rules and business models and sets of other users or, additionally, content — minus key metadata, for example, can be shared through an intermediary layer, with users on other social networks.
This can work technically — although it’ll be much harder to execute than it is to summarize. It can be funded via interconnect fees analogous to those paid by telcos worldwide, but a robust interconnect business model will be also challenging to build.
The demonstrable neutrality of an interconnect abstraction layer is essential, and it’s possible that it’d require some decentralized technologies, to assure the integrity and autonomy of the interconnect, and to make it resilient to challenge.
But will Facebook and WeChat play? Will they allow their users to push content through an API? No. No. No.
Not unless: alternative service providers are already using it and enabling value at scale or on a path to scale. Not unless: some government agencies (The EU, for example, or California and New York) require it. Not unless: critical masses of users start using intermediary layers.
There are, today, signs that the creation of multiple interconnect services might enable this to emerge, at scale, soon. Here are three:
Three emerging examples
Twitter BlueSky
Not much is known about this new initiative, other than the few indications given during its December 2019 launch.
The stated goal of the small investigation team is “to make disparate social media networks more like email, so that users could join different networks but still communicate with each other no matter which one they’re using.
Shared technical standards would also make it easier for users to gain some control over how these networks recommend content, which could reduce the tendency to guide users to the most outrageous material and users in hopes of keeping them engaged. It could also make it easier for the social networks to enforce restrictions against hate speech and other abuse, essentially helping them share the load at a lower cost.”
ActivityPub
The best-known (still little-known, tbh) project aimed at interconnecting social media is ActivityPub, which summarizes its goals as follows:
- ActivityPub protocol is a decentralized social networking protocol based upon the ActivityStreams 2.0 data format. It provides a client to server API for creating, updating and deleting content, as well as a federated server to server API for delivering notifications and content.
Mastodon is built using ActivityPub, using “… ActivityPub protocol to enable Mastodon servers to talk to each other; that’s the basis of the “federation” we also like to bring up. Federation is what you already know from e-mail, even if you may not know it by name: It’s the concept of servers hosting users that can talk to users from other servers. That protocol pins down on paper how exactly such inter-server communication would look like, using a vocabulary that can be applied for a variety of purposes.”
ActivityPub’s published object description is relatively arcane, and intended for expert code users — thus, the implementers of possible systems to interconnect social networks through an intermediate software system. In its self description, ActivityPub enables users in a mature implementation to …
- POST to someone’s inbox to send them a message (server-to-server / federation only… this is federation!)
- GET from your inbox to read your latest messages (client-to-server; this is like reading your social network stream)
- POST to your outbox to send messages to the world (client-to-server)
- GET from someone’s outbox to see what messages they’ve posted (or at least the ones you’re authorized to see). (client-to-server and/or server-to-server)
Digi.me
This project, based in Britain, was originally focused on enabling users’ to have full agency over their personal data, starting with social data, then adding health, finance and other data sources. It retains a B2C or B2B focus, building a platform for others to use.
Its website proclaims: “We’re enabling a decentralised world where people can control and benefit from their data with peace of mind that their privacy is respected” and “We embrace privacy by design principles and are working to make a decentralised world reality. Our distributed architecture and business model ensure we are never able to touch, hold or see user data — or tell people how or where to use their data.”
Digi.me aims for its data intermediary platform to be broadly used. It is the underpinning of the UBDI project — which aims to use personal data as the basis of income. UBDI describes its goals as “Giving people an income & an equity stake in the future value of their data” UBDI captures user data on spending, etc., and these data are shared through Digi.me’s anonymizing platform so that customers get the statistically significant data and patterns they need, while preserving privacy of users — and letting users share in the revenue of the market research projects.
Last words
I don’t foresee Facebook’s collapse, but I do look forward to a time when I can open Facebook, or some other media app, and see there fresh, interesting, poignant, intriguing posts, broadly available messages and private communications, and cat videos, from around the world — specifically including from my friends who aren’t using and perhaps aren’t able to use Facebook.
Other notes
This piece was originally posted on 6th January, 2020, and has been occasionally updated since (including the links below).
Other reading:
Jay Graber’s excellent note comparing federated and peer-to-peer network approaches.
Further writing on what a decentralized Twitter could look like from Marvin Ammori
Protocols, not platforms, in which Mike Masnick observes that much thinking on what is wrong and what can be achieved with social media is constrained by a reliance on the capabilities of the protocols that shape current media. Build new ones, with different capabilities for different outcomes!
My own gentle prior note about the reasons offered for decentralizing.
A Google docs version of this document, including full footnotes, with links, citations, and snarky comments.