You're viewing a single thread.
I just hope that it will be more distributed than Matrix and not everyone registers on lemmy.ml (matrix.org in case of matrix) so the decentralization works for real here instead of 90% (exaggerating, don't know the numbers) of the user base sitting on one instance :)
with how much .ml is struggling to handle the load I would've expected even more to pick different instances, though the situation seems better than how mastodon.social was during the twitter migration
the upcoming centralization issue sounds like it'll be the communities themselves all being hosted on .ml, not accounts. can't want to see how that one is gonna play out
they need the communities to be able to sync their posts/mod-team with other communities. That way communities aren't dependent on one site.
Yeah this is my concern with lemmy overall. Like if a server shutsdown, as far as my understanding goes, the communities and accounts go with it. Like yeah decentralized is great for democracy purposes but the hard line separation makes it a hard ask on time investment. And I'm sure it's more likely to happen to larger instances than smaller since cost will be the hugest factor. I'm sure remaking your account isn't a big deal but we've seen with Reddit a sub/community can be irreplaceable at times.
The first time a handful of large subs are lost on Lemmy I'm sure it'll have great affect on how the community views Lemmy.
Yeah definitely. There needs to be migration options between instances if federation is going to work.
This is probably a big ask from users but I wonder if something can be implemented into the protocol to create a bandwidth pool of sorts to host the various instances. Something similar to how Siacoin operated. If I recall correctly, users could install software on their computers that encrypted a chosen percentage of their drive, then used their internet connection to tie it into a decentralized file hosting network. (and rewarding the user with their currency - though that's not my focus here)
Files uploaded to the hosting platform would have redundancies created and uploaded to multiple of these hosting nodes to ensure reliable availability.
It'd be interesting to have a similar concept for the fediverse so we don't have issues like a loss of users and content if an instance goes down.
I just signed up for my local one, I see that is lemmy.ml.
That'll do for now. I will look around for alternatives once I got used to the whole federated thing more.