[sanderjd]

I'm with you here. I'll be bullish on federation when I understand what non-ideological problems it solves. I totally share the ideology! (And I get the sense that you do too.) But that's not enough. It needs to be useful to be durable.

[goodpoint]

Claiming that federation is ideological is itself a very ideological statement.

Being able to have an independent platform instead of a corporate-controlled one is very useful in practice.

[sanderjd]

> Claiming that federation is ideological is itself a very ideological statement.

No, it isn't.

Federation is ideological because it is not (yet?) in and of itself a useful feature. It is a feature desired by people (like myself) who have an ideology that attracts them to the idea.

Ask yourself: What is the selling point of federation / decentralization to someone who does have any ideology whatsoever with respect to technology? I want there to be a good answer to this question, but I've yet to see a compelling one.

[mavhc]

When the site you use goes bad/offline/whatever, you can easily move to another. If you're moving from a site that went bad/offline/whatever, do you want to lose all your data again next time?

[the_third_wave]

Federation is what keeps SMTP in place as the default electronic communications channel. Alternatives come and go but SMTP abides, there to cover your posterior when that fancy instant shout channel app flips over and goes the way of the dodo.

While many of the current proponents of federation are wont to explain the concept in line with their own ideologies - usually in the form of references to anti-Capitalism, Anarchism or some form of Marxism, often mixed in with a heavy dose of MDS when it comes to proposing alternatives to Twitter - this does not make the concept of federation ideological. It just makes it a widely applicable concept which has shown its usefulness for a very long time.

[sanderjd]

I don't think people fall back to SMTP when a chat product goes away, I think they find a new centralized proprietary chat product.

SMTP survives in its federated form because of path dependency. And people didn't start using email because it was federated, they started using it because it was an entirely new capability to most people. The federation was how its creators managed to implement such a thing at that period of time, but nobody in the 90s was saying "email is great because it is decentralized!", they were saying "email is great because I don't have to wait days for a piece of paper to travel across the continent and it's cheaper and faster to type than to make a phone call!".

Another commenter put this well: decentralization / federation of social networks thus far seems to be a feature but not a benefit. What is the benefit to users, what can they do with a decentralized network that they couldn't otherwise do? So far, the answers to this are ideological, the benefit is "I don't want to use an application centrally run by XYZ corporation, and this lets me achieve that goal". I think that's incredibly reasonable, but it is downstream of ideology rather than utility.

[the_third_wave]

> I don't think people fall back to SMTP when a chat product goes away, I think they find a new centralized proprietary chat product.

...after they told their "friends" (i.e. other users of the failed app) via email about that hot new app that is totally going to be the place to be. That is what I mean with "SMTP abides", it is there and will be there while the centralised proprietary churn comes and goes. SMTP is easy to set up, compatibility is close to guaranteed - Google's attempts to turn Ee-mail into Gee-mail have failed, Microsoft never managed to extend/extinguish it - and it runs on just about every piece of hardware known to mankind. Upkeep is simple as well, the spam problem has been solved a long time ago, email generally "just works".

It can even work as a chat server by using something like Delta Chat [1] if you're turned off by the "old school" user agents.

[1] https://delta.chat/en/

[sanderjd]

> ...after they told their "friends" (i.e. other users of the failed app) via email about that hot new app that is totally going to be the place to be.

There may have been a time when people spread the "hot new app" via email, but it ended decades ago at the latest, when Facebook came out. But I dunno, even before that, I heard about this kind of thing on AIM or one of the many other ones that already existed at that time.

But you don't have to convince me that email is great! Or that federation in general is architecturally satisfying. I'm right there with you and most other tech enthusiast.

But that's all beside the point being made here. The point is that federation is not why email was successful, or why it persists. It was just an implementation detail of a capability that was fundamentally novel to most people when they first came across it.

But that novelty is not the case for "Twitter - but decentralized!" or "Goodreads - but decentralized!".

[the_third_wave]

> But that's all beside the point being made here. The point is that federation is not why email was successful, or why it persists. It was just an implementation detail of a capability that was fundamentally novel to most people when they first came across it.

Nope, hard disagree there. If the mail-related protocols had not been federated - i.e. had email relied on a single-source centralised server - it would not even have survived into the 90's. It would have come up against a host of commercial competitors and eventually would have succumbed to some "Microsoft Network"-like thing which would have come pre-installed on Windows 95. Only Greybeards would use email, the latte sipping set would use Apple Mail, most of the rest would use whatever Microsoft presented them with. That was the original intent of the Microsoft Network as described in the first (hardcover) edition of Gates' book "The Road Ahead" [1] in which he envisaged a future where the internet gave way to the "Information Superhighway" based around proprietary technologies.

[1] https://en.wikipedia.org/wiki/The_Road_Ahead_%28Gates_book%2...