Email is an example of how federation fails, not how it works. It is now centralized in the hands of a few providers with the power to blackhole independent actors.
That centralization mostly occurred due to the economics of below-cost pricing though. It's not necessarily due to federation itself failing as a mechanism for email.
In other words, Gmail is the most popular email provider because it's in Google's interest to have people be logged in all the time so they can personalize search results, so that incentivizes them to make a very good email system and then give it away for free. Giving something good away for free because you gain indirect benefits via some other business will rapidly centralize more or less anything, which is why there are at least theoretically rules against market dumping and tying (which aren't really enforced in the software world, but that's another matter).
We can imagine a parallel universe in which search is far more competitive, with lower margins, and thus Google couldn't financially justify subsidizing consumer Gmail for so many years. In such a world it's likely that there'd be more players, perhaps they wouldn't be as good but there'd likely be more of them, there'd probably be companies that specialized in selling spam filtering tech to them and it'd look more federated than what we have today.
Freedom doesn't mean the requirement to spend the community's resources propagating material that the community agrees is harmful or objectionable. That would be the opposite of freedom.
The ability to refuse actors that the community doesn't want to spend resources on is essential to freedom.
Every actor has the freedom to choose whether to conform to community norms, or to form their own communities with different norms.
This argument comes with the rather absurd assumption that the only material that's being blocked is that which the community agrees is harmful. When more often than not it's simply the material that is from individually small enough third-parties where the content doesn't really matter.
On top of that, there isn't community agreement as much as there is community ignorance. For every case we hear about of someone being banned or blocked by an automated system despite doing nothing wrong, there are hundreds more similar cases which simply don't gain enough traction for the 'community' to hold the service provider responsible.
> This argument comes with the rather absurd assumption that the only material that's being blocked is that which the community agrees is harmful.
The new orthodoxy requires us to assume that all censorship is in good faith. "What were you doing to make them censor you?" is the question of the day.
> rather absurd assumption that the only material that's being blocked ...
It's fallacious reasoning to assert that any system that is less than 100% perfect should be changed.
If the current system meets the community's expectations 98% of the time, and alternative systems meet the community's expectations 90% of the time or less, it would be silly to throw out the current system.
That's like saying stop signs at intersections should be removed because sometimes I have to stop when there are no other cars there, and sometimes other drivers fail to stop when they should, and some stop signs have been defaced or knocked down.
This point is pointless. The question is whether federation works, or whether it instead degrades into a short list of large providers.
It's amusing that the first response to that point being made is to miss it, and to somehow paint the isolation of smaller servers as a strength of federation (because it gives users control over what they see.) Then the next one is literally an argument that decentralization will inevitably centralize, because that centralization will satisfy the communities expectations 98% of the time.
The last paragraph wanders very close to claiming that it is childish and entitled to expect your decentralization not to be centralized. Doesn't that just enthusiastically reinforce idlewords's claim?
So, if certain minorities have more representation in crime statistics in an area, it is acceptable to have a bias in random searches towards them because doing so meets the expectations of the majority to have to deal with neither crime nor random searches compared to the alternative of unbiased searches which lower arrest rates and expose the majority to more searches?
Even putting aside the slightly absurd difference in stakes with my example, there's also the issue that you're just arguing for centralization, which defeats the point of a system designed for federation in the first place.
Yes, that's the universal argument that can be used to argue for or against anything, depending on what you find fashionable at the moment.
The concrete truth is: majority of users will do whatever to make their e-mail (or other form of communication) work, so the easy solution to spam is to blackhole anyone that isn't coming from or vouched by a major provider, and have the minority with specialized needs jump through hoops to get unblocked (or preferably, to change their needs/preferences so that mainstream provider is sufficient).
There isn't any actual user community here. GMail address owners aren't a community. Twitter users aren't a community. Mastodon users aren't a community. The only community is that of the federation instance owners, and it operates as oligarchy. The major players decide who they'll peer with, the minor players have to accept it if they want to be peered with. The actual, non-crazy-activist users, have no power, and are forced to either use the "blessed" providers and accept their rules, or not use the service.
This is already the case for Mastodon for quite a while, BTW.
Freedom for titanic companies like Google to control what can or can't be said through enforcement on their monopolistic platforms is also not freedom for the individual.
Your idea here works when the system is fragmented enough that these choices are personal choices, but the existence of so many large platforms makes their choices systemic, not personal.
I'm skeptical of everything, and have been a long-time skeptic of Mastodon in particular, but I'm less concerned about this particular argument.
* People don't need nu-Twitter to be perfectly reliable they way they do for mail
* For that matter, nu-Twitter doesn't really need to be a unified namespace; it can be less coherent even than email addressing and still add value, the way RSS feeds did.
* In the years before Google centralized all of email, it was pretty common for people to have multiple email identities on different services, which seems like the worst this is going to get.
I don't love the software, and the user experience definitely isn't ready for prime time yet, but the basic model seems workable?
I agree. Did RSS feeds suffer spam? I don't think so, and the reason was because you have to actually follow someone to receive things from them (whereas with email you get to receive everything anyone sends you so long as they have your address)
It's still viable to use a minority email provider, even if that now has to be a medium-size ISP (at minimum) instead of your own server. There's still a difference between no choice and many choices.
