Partly. Except the time different Mastodon instances were not federated much or at all. If you wanted to go follow someone on Mastodon you had to know the exact server they were on. In an environment like Reddit and Lemmy where you’re there for the communities instead of the people that isn’t an issue. But if you want to go follow some specific podcaster you need to know the instance because there’s no guarantee that whatever instance you happen upon is going to be joined up with the one there on.
Everyone was busy running their own servers and not trying to tie everything together. It was a thing that could be done but a thing not enough were doing.
That’s nonsense. I’m on one of the main servers, and like 90% of my feed is from other servers, and it includes lots of small servers. And that’s been true for years.
It’s try the search function was bad prior to earlier this year, but it’s improved a bit. And if you are looking for someone specific, then presumably their account would be listed somewhere on their website?
That sounds worse than I thought it was. I just assumed Mastodon was like Lemmy, where every instance federates with every other instance basically by default and there’s only some high-profile defed exceptions.
A Fediverse where federations are opt-in instead of opt-out sounds like actual hell. Yeah, more control to instances, hooray, but far less seamless usability for people. The only people you will attract with that model are the ones who think having upwards of seven alts for being in seven different communities isn’t remotely strange or cumbersome. That, and/or self-hosting your own individual instances. Neither of these describe the behavior of the vast majority of Internet users who want to sign up on a platform that just works with one account that can see and interact with everything.
Mastodon federation is not opt-in. As soon as anyone on one server is following one person on the other server, the servers are fully federated. From there, it’s opt-out, via blocking.
I just assumed Mastodon was like Lemmy, where every instance federates with every other instance basically by default and there’s only some high-profile defed exceptions.
That’s…Not how Lemmy works either. In fact, and someone may correct me if I’m mistaken here, your hell is sort of how it works as I understand it. Instances don’t have any built-in crawlers to seek out others running on ActivityPub with the same software, e.g. Lemmy or Mastodon or the like. That’s genuinely been one of the biggest stumbling blocks with the whole protocol, as discovery is largely a manual affair. The only crawlers we have are the people using the service and following remote people or communities or channels from other instances to let the one we’re on see them.
One of the basic reasons for this that I’ve read is that it’s related to handling scaling, as each instance trying to handle all of the data of all the people on each other instance right away would bog down the servers and probably crash them. It also arguably works out, to a degree, that there’s a good chance not everyone on each instance is of interest to each other anyway, so you may not want or need each server to know about every other server’s people/channels/communities/etc.
But I’m going to stop before I get too much further into the weeds of all this. The irony is that the simplest solution to discovery issues with all of this presently is to invite those you want to have a similar experience to you, or want to connect to with the fewest jumps, to the same instance as you to mitigate any of those issues. Does that tend to undermine many of the benefits of it all? In a lot of ways, yeah, but that’s where many ActivityPub platforms are at currently, at least the more popular ones as I understand them.
Partly. Except the time different Mastodon instances were not federated much or at all. If you wanted to go follow someone on Mastodon you had to know the exact server they were on. In an environment like Reddit and Lemmy where you’re there for the communities instead of the people that isn’t an issue. But if you want to go follow some specific podcaster you need to know the instance because there’s no guarantee that whatever instance you happen upon is going to be joined up with the one there on.
Everyone was busy running their own servers and not trying to tie everything together. It was a thing that could be done but a thing not enough were doing.
That’s nonsense. I’m on one of the main servers, and like 90% of my feed is from other servers, and it includes lots of small servers. And that’s been true for years.
It’s try the search function was bad prior to earlier this year, but it’s improved a bit. And if you are looking for someone specific, then presumably their account would be listed somewhere on their website?
That sounds worse than I thought it was. I just assumed Mastodon was like Lemmy, where every instance federates with every other instance basically by default and there’s only some high-profile defed exceptions.
A Fediverse where federations are opt-in instead of opt-out sounds like actual hell. Yeah, more control to instances, hooray, but far less seamless usability for people. The only people you will attract with that model are the ones who think having upwards of seven alts for being in seven different communities isn’t remotely strange or cumbersome. That, and/or self-hosting your own individual instances. Neither of these describe the behavior of the vast majority of Internet users who want to sign up on a platform that just works with one account that can see and interact with everything.
Mastodon federation is not opt-in. As soon as anyone on one server is following one person on the other server, the servers are fully federated. From there, it’s opt-out, via blocking.
That’s…Not how Lemmy works either. In fact, and someone may correct me if I’m mistaken here, your hell is sort of how it works as I understand it. Instances don’t have any built-in crawlers to seek out others running on ActivityPub with the same software, e.g. Lemmy or Mastodon or the like. That’s genuinely been one of the biggest stumbling blocks with the whole protocol, as discovery is largely a manual affair. The only crawlers we have are the people using the service and following remote people or communities or channels from other instances to let the one we’re on see them.
One of the basic reasons for this that I’ve read is that it’s related to handling scaling, as each instance trying to handle all of the data of all the people on each other instance right away would bog down the servers and probably crash them. It also arguably works out, to a degree, that there’s a good chance not everyone on each instance is of interest to each other anyway, so you may not want or need each server to know about every other server’s people/channels/communities/etc.
But I’m going to stop before I get too much further into the weeds of all this. The irony is that the simplest solution to discovery issues with all of this presently is to invite those you want to have a similar experience to you, or want to connect to with the fewest jumps, to the same instance as you to mitigate any of those issues. Does that tend to undermine many of the benefits of it all? In a lot of ways, yeah, but that’s where many ActivityPub platforms are at currently, at least the more popular ones as I understand them.
It sounds like an asocial network - something for me finally, perhaps :)