Clarifying the origin and evolution of the term "the fediverse"

Continuing the discussion from SocialWebFoundation - what do people think?, where this topic was discussed at some length (and I would have been wiser to fork it off into its own topic much soon, sorry everyone!)

Well no, and if you think that’s what anyone’s arguing, I think you’ve got the wrong end of the stick. Although it’s possible Evan and SWF see it this way, in which case I would disagree with them. But I think their perspective is more W3C-centric. Something like;

  • the web is what W3C standards define it to be.
  • The social web is what the social subset of W3C standards define it to be
  • The existing fediverse adopted one of those standards (well, the server-to-server part), making it part of the social web.

(Is this a fair summary @eprodrom?)

To me, the fediverse is a diverse network of people, unified by a shared approach to technology. Currently that technology is unified by implementing AP, although previously it was OStatus, and in future it could be something else.

AFAIK the Pump API was only ever used by pump.io (the app hosted at identi.ca), just as DFRN was only used by Friendica. The fediverse.party definition only includes protocols that have been used for interop between two or more fediverse apps; OStatus, Diaspora, Zot, and AP. But given that except for Diaspora, all the apps that ever used those protocols started using AP within a year of the final spec being published (except for a few that died before they could), it’s not inaccurate to say that the fediverse was multiprotocol. But we made an intentional decision, as a social movement, to converge on one protocol.

I agree with this up to a point. As with “Open Source”, there’s no central organisation that owns a trademark on “fediverse”, and can dictate an “official” definition. But like “Open Source” it has an underlying philosophy, which leads to a broadly agreed definition that definitely includes some things, and definitely excludes others. Like “Open Source”, people with agendas regularly try to broaden the definition into meaninglessness, in pursuit of their own financial or ideological agendas. So people who care about its history and shared philosophy have to push back.

I trace the origins of the fediverse back to a social movement that began in the mid-noughties, where some of us started organizing around concerns that DataFarming was taking over the net. Producing things like autonomo.us, and the Franklin St Statement. And events like the first IndieWebCamp, and the Federated Social Web Summit in 2010. The pioneers of today’s fediverse were all, loosely speaking, part of that social movement.

In the noughties, such concerns were mostly ignored, or perhaps laughed at, as were online privacy concerns in general (“tin foil hat” etc). But this began to change with the Snowden disclosures in 2010, and these concerns are so mainstream now there’s even a buzzphrase for it; the techlash. Our social movement, in cooperation with many others, has succeeded in getting society to notice it has a DataFarming problem.

As a result, the 2010s produced a plethora of software and services pitched as pro-privacy, including dozens of decentralised social network projects. As well as the Matrix project and few others that are still around, this included the first fediverse apps; Laconica/ OStatus (which became StatusNet, and now GNU social), MistPark (which became Friendica/ DFRN, which evolved into Hubzilla/ Zot), and Diaspora. Side note: Many people forget that the original Diaspora pitch was to create a P2P app with no servers, the pivot to a federated model (using a fork of the OStatus protocol) came as a surprise to many crowdfunding backers.

Then the wider dissemination of the BitCoin paper kicked off the “Web3” wave of distributed social network projects. Now there were even more bleeding edge social network projects competing for mindshare.

Now all of this is a technologist’s dream. So much fun new tech to play with! But assessed against our social movements goal of creating viable replacements for the DataFarms, it falls short. That’s because it fragments the network effects of each option, and leaves people scratching their head about which software is production ready, and just as importantly, which one they need to be where the people are.

Imagine if there were dozens of mail federations using incompatible protocols, and you had to figure out which one the people you wanted to talk to were using. Then which software supported it, and which of that was production ready. Email could never have become the common carrier it is, to the point of replacing snailmail for anything that can be digitised.

So because our social movement’s goal was to replace the DataFarms with a protocol-based common carrier, we started talking about how to get as many of the apps as possible using one decentralised protocol. On the existing networks, in Loomio groups, in dev forums like SocialHub, etc. The formal expression of this was the W3C process that created AP. But there’s no way all the apps using OStatus, Zot, and Diaspora protocol (except Diaspora themselves) would have all implemented AP within about a year of the spec being published, if there wasn’t a pre-existing consensus on moving to a single protocol.

Now I don’t know anyone passionate about the fediverse who says people shouldn’t develop on other federated protocols or use software that does. We all use email, and most of us also use XMPP or Matrix apps (or both), and experiment with anything else we come across. But we tend to argue for bringing more apps into the existing fediverse, and against fragmenting it across multiple incompatible protocols. Because this is opposed to our goal of creating a social network common carrier equivalent to email, and we’ve done a lot of work to progress as far as we have towards that (witness the mainstreaming taking place since the inflection point of Eternal November).

This is not the nature of our disagreement at all, as I hope the above makes clear. The actual difference is protocols vs. people. You define the fediverse from a top-down perspective, in terms of the protocols you think it should include. I define it from a bottom-up perspective, based on the protocols the people involved have chosen to develop and use. You’ve come in as an outsider to bring the multi-protocol gospel to the savages. I’m here as an insider (since the original StatusNet iteration of identi.ca), politely explaining that we already tried that, and then found that using one protocol worked better for us.

Quite the opposite. I want it to be like Open Source, a living definition, based on the very specific history and philosophy of the social movement that coined the term and did a tremendous amount of (mostly unpaid) work to bring the current network together.

If you want to talk about a network tied to AP, come hell or high water, I would vastly prefer you call it the Activityverse rather than the fediverse. But if I suggested that I was going to use the name Steve Bate for a broader purpose that reflects my agenda, and that you ought to pick another name, you would be well within your rights to tell me to move along. Well the social movement that created the fediverse is not going anywhere, and if you don’t like how we use the term we created, you are more than welcome to move along.

I can’t speak for them, but I suspect their goal is to facilitate the evolution of the W3C standard called ActivityPub to address the missing stairs that led to the creation of ATProto and Nostr (eg there are already FEPs for portable accounts thanks to @macgirvin’s work). Leading to a future where all the apps currently being developed for the fediverse, the ATmosphere and the Nostr-verse end up in one network, using a protocol called ActivityPub. But which may or may not bear any resemblance to ActivityPub in its current form.

The web has followed a similar path. Standards like HTTP and HTML have evolved significantly, to keep up with the needs of implementers. At times, splinter groups broke off to develop non-compatible version to address missing stairs. But eventually standardisation processes reunited the camps, and we still call the web protocols HTTP and HTML.

In the context of all of the above, I’m yet to be convinced that’s a bad thing. Since many of the people opposing them have the opposite goal of fragmenting the fediverse, and the social web in general, into multiple incompatible networks, that only makes me more inclined to support them.

Always a pleasure. Since you’re taking an interest in the history (unlike some), I’ve been keeping track of coverage of the fediverse here;

It’s a bit out of date, especially given the events of the last few days. If I get time, I’ll do an update, and links to add are welcome. Here, or in a fediverse.party issue.