I've been wondering about this for the past week and given the trouble Lemmy users have had with the Nicole spam, would it make sense for large Lemmy instances to switch to a whitelist approach for federation?
Instead of automatically federating with every new instance, what if we set up a system where federation had to be requested and approved? New instances could submit a request to federate, and the panel of federated instances could evaluate it before accepting. Any instance added to the white list would ideally be whitelisted across participating instances.
It might help with moderation challenges, reduce spam and bad actors, and give communities more control over the content that appears. BUT it adds unnecessary friction and turns the Lemmyverse into a closed space which goes against the idea of federation.
Curious what other people think and how we can brainstorm an approach for this kind of moderation issue long term.
Edit: please do participate in this conversation instead of downvoting ?
Maybe something like a time delay, or thresholds of some kind could be added into this equation?
Say, theoretically, everyone switches over to a verified whitelist model, instead of an automatic approval of any new instances.
Now, I don't think you're gonna get some kind of... UN of large instances to collectively do some kind of formal process that is shared between all members, and then somehow their collective decisions are binding for all members of this... instance UN.
But, individual instances could adopt and manage their own less permissive whitelist model, with their own vetting processes.
So an example could be:
Don't auto federate to every new instance.
Instead, wait for the instance to be... 2, 4, 6 weeks old, wait for it to have so many unique users, so many daily posts or comments in whatever timeframe... then when that threshold is passed, it gets auto whitelisted.
If a new instance owner wants to accelerate federation with your instance, new instance owner can send out some kind message describing their instance, overriding this process.
I am not 100% sure how the actual technical backend here works, how easy or difficult this would be to implement, how much workload this would put on instance admins.
If anything like this is possible, I really think you'd want to keep the 'sovereignty' of how the process would work still within the sole control of each instance, and not just... make a super-instance governing body that makes binding decisions for all accepted members.
That would be a mess, and would... just make the 'panel of big important instances' into basically the UN security council.
But at a per instance level, it doesn't seem unreasonable that individual instances could consider their own sets of federation criteria other than just 'automatically accept everything and only blacklist after massive drama or a botnet instance arises'.
Yeah i dont think any from of UN like organisation between instance is even feasible without opening a can of worm when it comes to who has authority over the Lemmyverse, your ideas are a lot more thoughtful and elegant