3 ms·
> These are talked about as problems, rather than being the primary selling point. The Mastodon implementation of this is legitimately flawed. The issue is yo
by AnthonyMouse 4y ago
> These are talked about as problems, rather than being the primary selling point.
The Mastodon implementation of this is legitimately flawed.
The issue is you need users to be independent of servers.
That isn't hard to implement. Make the user ID in the protocol the user's public key and user@domain is just a human-friendly pointer. If you transition from user@a to user@b, your pointer changes -- or you can use both for an indeterminate amount of time -- but the public key stays the same and so everyone who followed you or signed up for your feed still sees it.
The current implementation is broken for multiple reasons.
First, it makes you have to care who is hosting you, because if they become adversarial or go out of business or you become large enough to want to self-host, you lose your account. That shouldn't happen.
Second, with users tied to hosts, it becomes "convenient" for hosts ban entire other hosts, which is poison because it's the dynamic that led to the bulk of email moving to GMail and Outlook. Once providers are big enough they can ban smaller ones ("for spam filtering") to exclude competitors. The protocol needs to prevent this by making the host fungible. Not doing so is a significant design flaw.
- midoridensha 4y agoWell, it's too late now, so I guess we're stuck with it. Just look at email for example: its fundamental design is quite flawed as well, since it was conceived before anyone dreamed of spam, but now there's no way to fix it aside from hacked-on additions like spam filtering. The protocol, devised in the 1970s, can't be changed now, and we can't get everyone to upgrade to a new protocol. Why didn't the architects of Mastodon think of this stuff before? It should have been pretty obvious that people might want to move their accounts between servers. But now, they'll *never* be able to. The only way it can happen is if someone creates an entirely different social network like Mastodon and somehow gets everyone to switch to that instead, but it's probably too late now.
- AnthonyMouse 4y agoMastodon still has like four users compared to any of the major social networks. Fix it now even if it takes a breaking change before it hurts more later, or do better with whatever thing actually does eventually rival the user count of Facebook or email. > The protocol, devised in the 1970s, can't be changed now, and we can't get everyone to upgrade to a new protocol. This is untrue. There is a clear path to adding things like DKIM or SPF. First you add support to the major servers and publish a spec, then you give everyone else a reasonable amount of time to implement it, then you turn it on. This breaks anyone who hasn't implemented it yet, but by then that's only people who suck, and anyway so does dumping all of the mail they send to spam with nothing they can do about it. The problem with email is that the major providers don't want to do more things like that (e.g. implement hashcash or similar), because it's better for them for the smaller providers to be destroyed than to design a protocol improvement that facilitates their existence. Which is why it's better for it to be there from the start. But the best time to do it was yesterday and the second best time is today.
- midoridensha 4y ago>Mastodon still has like four users compared to any of the major social networks. Fix it now even if it takes a breaking change before it hurts more later, It's already too late. It should have been fixed before Twitter nose-dived. Coming up with a replacement at this point isn't going to work: the Mastodon name is already out there, and this is the time when an alternative has to be ready and usable: it's first-mover advantage. >There is a clear path to adding things like DKIM or SPF. First you... So why hasn't this been done yet? We've had decades. >The problem with email is that the major providers don't want to do more things like that Exactly my point. Email *can't* be fixed at this point because of first-mover advantage. Incumbents don't want to fix it. Users moved to email because it was available when they were ready, and it was the best alternative at the time. Mastodon is likely in the same place now. Something better doesn't exist yet, and doesn't have any name recognition. Mastodon itself can't be fixed because the designers like it how it is: if they wanted a better design, they would have designed it that way from the outset. So if Mastodon catches on, we're stuck with it as-is.
- AnthonyMouse 4y ago> It's already too late. It should have been fixed before Twitter nose-dived. There's nothing existential wrong with Twitter. Elon Musk generates drama which the media plays up for clicks. It's an open question how many of the people creating Mastodon accounts because a reporter mentioned it are going to be actively using them in six months. And if it continues to be the same kind of garbage fire subject to the whims of oligarchs as the networks people are fleeing, that's not likely to be many. Fixing this kind of problem is how you get people to stay. > Coming up with a replacement at this point isn't going to work: the Mastodon name is already out there, and this is the time when an alternative has to be ready and usable: it's first-mover advantage. Twitter has first-mover advantage. Mastodon is still a tiny thing. > So why hasn't this been done yet? We've had decades. DKIM and SPF were done. They're widely supported and work as designed. > Incumbents don't want to fix it. This is an obstacle, not a death ray. Find a way to exert pressure on the incumbents to fix the problem. Isn't this supposed to be the benefits of centralization I keep hearing about? There are only a couple of parties you have to convince in practice to get something changed? > Mastodon is likely in the same place now. Something better doesn't exist yet, and doesn't have any name recognition. There will be a fresh omnishambles causing people to seek alternatives a year from now after you've implemented a proper protocol for people to switch to. Do it now so you're ready for then. > Mastodon itself can't be fixed because the designers like it how it is: if they wanted a better design, they would have designed it that way from the outset. That doesn't follow. People frequently make mistakes and regret them later. > So if Mastodon catches on, we're stuck with it as-is. So don't use it until they fix it. It's not like Twitter was actually incinerated. What good is switching to a known-broken technology? Wait for the one that works, and in the meantime build it.
- nijave 4y agoEmail has undergone enhancements over the years that have widespread support. IMAP, DKIM, and SPF all come to mind.