I was reading this post by Alyssa Rosenzweig and found it to be a very enjoyable read.
https://rosenzweig.io/blog/the-federation-fallacy.html
Rosenzweig tells us that although we have created federated social media, we should expect them to centralize. That is what has happened to email and empirical data demonstrates that this is happening for mastodon as well.
To be clear, this post was made in march 2019 and I am not aware whether she has changed her opinion. The current state of decentralization within the mastodon ecosystem may also have changed.
I will challenge the statement that email and mastodon has an equal potential to decentralize the web. Which is not explicitly said in the post, but which I feel is a sentiment.
I will talk about mastodon here for the sake of making the talk less abstract. This text will be applicable to other fediverse social infrastructures as well.
Introduction
We like to compare email and mastodon to showcase the flexibility that is lacking in centralized social media. This comparison is valuable because it makes these concepts much more understandable for newcomers.
However, treating email and mastodon as equal forces in the fight for decentralization is inappropriate.
Differences between email and mastodon
It is important to emphasize a core difference between email and mastodon.
The community element. Although email and mastodon are federated and decentralized, email doesn’t have a community element. Email only connects you to those you follow whereas mastodon is designed to connect you to your community and your federation.
How easy it is to self host is the selfhostability. Is it easy to install? Is it resource demanding? Is it possible to pay others to run the service? The better the selfhostability, the more people will be able to do it. It is essential for the decentralization potential.
Different mentalities
There are two mentalities. The globalization mentality and the community mentality.
Those with a globalization mentality view an instance just like an email provider. They want to subscribe to exactly the users they are interested in and that’s it. The community element doesn’t really matter to them because they don’t join an instance based on a theme. They could perhaps be interested in mastodon.social or mas.to.
Those with a community mentality will seek instances based on the community element of an instance. Those with a community mentality are more likely to subscribe to instances with strong themes. They could perhaps be interested in gensokyo.social, dolphin.town or meow.social.
Understanding the decentralization potential
To understand the decentralization potential for each of these solutions, we need to understand:
- who will feel compelled to host the solution.
- how will users elect their providers.
By doing this, we can determine the decentralization potential.
Email user
The email user has a globalization mentality and will judge an email provider by pricing, security and other technical criterias. Nobody choose email provider based on community.
Centralized actors has a much greater advantage of hosting when they only needs to focus on the technical aspects. They can push the prices down and all their spending can easily be advertised. Low risk, high reward.
The selfhostability of email is a technical nightmare or from what I’ve heard. You gain autonomy, but there is low community reward. High risk, low reward.
Based on this analysis, we should expect users to flock to centralized solutions.
Mastodon user
If the mastodon user has a globalization mentality, the users are most attracted to big instances. They have most likely been active on twitter and just interested in a replacement rather than actual structural change. Perhaps they view bigger instances as more sustainable. They are disinterested in thematic instances because that would arbitrarily limit speech.
If the mastodon user has a community mentality they will gravitate towards instances with a strong local or federated culture. They believe that as instances grows big, it stops being an interconnected community. They value small and thematic and value good moderation.
Mastodon can be very risky for centralized actors as it requires involvement in the community. Often times it can come at shareholders, electorate or userbase expense. We can see this all the time with various mastodon.social, twitter and facebook contraversies. It can conceivably be done right and in that case, the rewards could be great. But definitely not ideal for all centralized actors. High risk, variable reward.
How is mastodon’s selfhostability? You can allow others to run your instance through Masto Host or Spacebear Federation. And the code is designed to be easy to install. If operating systems such as yunohost properly supports the mastodon app, this will also increase the selfhostability.
So will Mastodon be a force for decentralization?
Mastodon could easily become centralized if enough people have the globalization mentality and selfhosting remains unaccessable/unused. Corporate actors can easily capitalize on the globalization mentality.
On the other side… If we can shape the masses into having community mentalities. If we can shape development to improve selfhostability. If we can create enthusiasm for digital freedom on the ground. The future will be decentralized.
CC BY 🧡
I also like to think of it this way. Even if Mastodon gets “centralized” we will at least not have a hard barrier forcing us to only use that centralized instance. If that centralized instance fucks up, there is at least always the possibility to create and join a new instance. Without the risk of losing contact with all the people on the old instance.
If Facebook fucks up, we can’t spin up a new facebook instance.
I don’t mind “soft centralization” as long as it allows for decentralization.
iirc, a large number of admins recently defederated from mastodon.social due to its lack of moderation. I haven’t noticed any major changes on my instance’s federated TL.
You’ve written a good analysis, OP. Yes, treating email and mastodon as equal forces in the fight for decentralization is inappropriate. However, email is not as one-sided as you have made it out to be:
- Email does have a community element: mailing lists may be the most obvious example, but email does have the ability to drive community discussion. What is holding it back is obviously its age and ease of use.
- Organizations use email: Many organizations will self-host email simply to manage mailing lists as well as email notifications. Having an email address at a domain signifies your membership in some sort of community; perhaps your workplace or school.
- Ubiquity:
- Email is still a more universal communication method than any other messaging protocol. You can probably reach anyone through email.
- There are many organizations, communities, etc. offering email services. They must allow their users to communicate with other servers: no email provider is large enough to disregard all the other providers.
- There is a widespread expectation that your email address will be able to reach any other email address. In this way it is the most federated, decentralized messaging protocol there is today.
I also have a point to make, both in reference to the blog post linked and your analysis: I would not dismiss federation because of some centralization of servers.
We have a case study in email to show us how a healthier (for all of its faults) federated system can survive without becoming wholly centralized. The best way to prevent centralization from dismantling federated systems is to build in an expectation that a server/service/offering in a federated system will always be able to communicate with another in the same system, like in email. As long as one federated server/service/offering/application does not become its own “brand” compared to the other servers and the base protocol/federated system is all the talk, all will be fine.
For example:
- Let’s chat with Element! vs Let’s chat with Matrix!
- Let’s join Lemmy.ml! vs Let’s join a Lemmy instance!
- What’s your Gmail? vs What’s your email?
Some of these are fair, but even 4 bigger instances is better than the current model of just one, like for twitter or FB. And the trend over time also seems to be towards more decentralization, not less. A project first starts out, and has a few instances that dominate, but more people join, and not only do these first instances grow, other ones do as well.
The author also seems to forget the original intention of the net, which is to have a completely distributed system where everyone is talking to each other via network calls… IE the server is a small as an app that every computer is running. Activitypub and matrix for example are just open specs / protocols; as soon as performant apps are made that can run on a single device, this will be a reality. Eventually matrix will get there, they’ve been doing testing with dendrite running as WASM in a browser that already works. Lemmy might be there already tbh, if we didn’t rely on activitypub’s current need for domain names.
Activitypub and matrix for example are just open specs / protocols
I think this is the most important observation of them all, and where the focus of decentralization should be: evolving these specs and the ecosystem that exists around them so that many different projects and app types find it ever easier to interoperate. I feel that all this focus on Mastodon only serves to distract from and obfuscate what our common goals should be: the fact that we should place our efforts on furthering the open standards and all of the technologies and projects that build upon them in their entirety. And that this constitutes a collaborative community effort.
The email user has a globalization mentality and will judge an email provider by pricing, security and other technical criterias.
And marketing or advocacy.
Is my post that uninteresting that people are going to ignore it completely and only talk about the referenced Rosenzweigs blog post instead?
By all means, my post could be uninteresting, not worth your time, unintelligible or simply wrong. But then I encourage you to refrain from upvoting or even downvote it.
If you really want to talk about the reference, you can do that here instead: https://lemmy.ml/post/84655
I just feel discouraged.
Literally the first thread under your post is talking to you, the OP, and addressing your points.
What were you looking for?
I posted my comment before that was posted. I was very happy to see it :)