idk it’s my first time (kind of) trying to do dataviz, I thought the original post was super interesting but like other comments said, it was hard to read.
I think in a way it’s mostly because defederation is complex to show as links, since it is the break of said link, but hopefully this type of graph is better ?
you do lose the “group” of instances view, though (and also I did not bother with per instance MAU as raw number + percentage, or sofware used)
let me know what you think!


It doesn’t even Include https://lemmy.today/ that has no defederation or blocking at all, and has never had, since the admin feels it’s up to the user and not the admin to decide these things. You know, free and open internet, remember?
Also one of the best themes of all Lemmy instances if you ask me.
Came here to say that!
I can’t comprehend why people would join instances in the list. If you don’t like a particular instance, just hop into the user settings and block it! Why let others set your blocklist?
because it’s less work.
Doesn’t take much normally. But still, I think Lemmy would surely benefit from adding custom filter lists that users can enable or disable at will.
You gotta ask the OOP about that, but my guess is, in a graph that shows instances that defederate, why include an instance that doesn’t?
Because it lets you see how common or uncommon defederations are relative to the total set of all instances (or at least, all minus the long tail of very-low-user-count ones). And although it would make the graph larger, it would also make it sparse and IMO that would also be easier to read.
It could also give a misleading impression that all instances defederate or are defederated by somebody, making people think it’s more common than it actually is. (Or maybe not doing that, if you include allish instances and it still ends up being not very sparse.)
I see your point I guess. Perhaps defederation is something that users want, and then they are happy to see who does it the most?
Im not like that myself, but I can see the point.