If fedi node A and node B both have an anti-spam rule, it makes good sense that when a moderator removes a post for spam that it would be removed from both nodes. But what about other cases? Lemmy is a bit blunt and nuance-lacking in this regard.
For example, the parent of this thread was censored despite not breaking any rules. More importantly, it breaks no rules on slrpnk.net. Yet the slrpnk version was also removed.
I’m not sure exactly what the fix is. But in principle an author should be able to ask a slrpnk admin to restore the post in the slrpnk version of that community, so long as no slrpnk rules are broken by the post.
It’s one thing for various nodes to federate based on having compatible side-wide rules, but they aren’t necessarily aligned 100% and there are also rogue moderators who apply a different set of rules than what’s prescribed for a community.
I only discovered the censorship today when trying to search for my own post to reference from another post and gain a history of the discussion. That’s another problem… the fact that censorship is so quiet. It’s not a shadowban, but if you don’t make the effort to check whether you have been censored, it has the same effect as shadow banning on reddit. Though the client should get the blame for that (and in this case I use the stock web app).
(edit) Important to know why shadow banning is such a bad idea: it fails to teach a user to be a better user. If you do not inform a user that they broke a rule, they don’t learn from their mistake. Users should receive a loud and clear signal that their post was blocked because it broke rule X, so they learn the rules and can avoid repeating the mistake. Lemmy needs to improve on this.
Indeed it makes sense from an efficiency standpoint. But there should be an override on every host.
Different venues have different rules, so the non-correctable status quo is not sensible. And in the case at hand, even the original post complied with the rules. This common problem is not limited to Reddit. Mods can-and-will abuse their power and apply non-existent rules. The mod log makes it optional to even specify a reason.
Different venues have different rules, but the moderators are enforcing the rules of their community, and as such the other venues should abide by that.
Say for instance I remove a post that’s overtly racist, but there’s some lemmy out there which is some Nazi white power bullshit, they shouldn’t be able to over-ride what is presented in my community just because it got copied to their server.
I’m not sure you understand that there are multiple copies of a message when multiple federated nodes are involved in a transaction. Every host involved has their own copy. They often delete their own copy not to censor but just because they are short on space. Every host admin should control their own persistent storage.
Software should serve its user. If you are developing a web browser, then the web browser should act in the interest of the user (the web “surfer”), not the website they are accessing. Developers sometimes get that backwards and design the client software to prioritize the interests of the server it connects to over that of the user. And this is what you’re suggesting.
If nazi instance is running Lemmy, the nazi admins are the user and their software should serve them. This is an important FOSS principle (“freedom 0”). In the non-FOSS world, sure, software serves the capitalist supplier. But in a FOSS context the Lemmy server should be working for the Lemmy admin (the user in this context) who runs it, regardless of their politics.
The tables can just as well be turned. A greenpeace instance might post something on a nazi instance which then gets censored on the nazi instance. Fair enough, but the greenpeace instance should have control over their own instance and be able to uncensor msgs on their own host.
(edit) It’s also important to understand the power balancing principles of the fedi. The fedi tries (though often fails) to balance out power. When a lousy mod takes a questionable action that then has a negative effect on other nodes, it’s a power imbalance. Ideally other nodes should not only have their own influence, but in a democratic sense it would be useful if nodes would become aware when other nodes overturn a removal, so they might consider their own intervention. Otherwise a one-off asshole can have too much influence.