There are bot instances cropping up that absolutely must be defederated in order to keep the platform viable.

Staying in federation with an instance that actively embraces bad actors increases the visibility of users here to those bad actors, and gives them access to our community. Defederating such an instance is a basic best practice in the Fediverse.

More importantly for those who wring their hands about not limiting the whole community -- failure to defederate from bad actor instances will be factored in when good productive instances with content folks here want to see decide whether to defederate us. (Remember that this place is already defederated by one prominent instance, which is a material detriment to users here.)

It is reasonable and normal to disagree about where the line is drawn in terms of what instances deserve defederation. It's often ambiguous what's a normal instance with sloppy moderation and a few bad apples[^1] versus what's a place that is run by and for bad actors.

There's a wide range of standards that can be applied. It seems like the general vibe can be broken down into three groups:

• Only defederate spammers and child porn
• Only defederate spammers child porn and tankies
• Defederate spammers child porn, tankies, and rampantly fascist troll farms

I don't think anyone has really advocated for anything aggressive than that on here (could be wrong)

[^1]: Although also important to remember that the point of the bad apples thing is that they spoil the whole batch if you don't take them out.

So what you're suggesting works in theory, but not in practice. There are times when simply ignoring bad actors or voting them down within a community is not enough to stop them from misbehaving. Try ostracizing a botnet. Or nicely asking trolls to stop posting racist memes. It won't work.

For the other instance's users they would still be able to see our communities and post replies in threads but only users from their own instance would see their messages so in the end the majority would think our communities are just dead and that would be the end of it.

Only a minority would go through the trouble of creating a second account just to troll and the few that would would be easier to moderate than a whole instance worth of users.

The experience on all other social medias show that it just doesn't work to try and coexist with extremists and a feat like askhistorians requires a team of mods with someone there pretty much 24/7 to watch what's going on and to automatically ban Holocaust deniers (for example)... And that was only one subreddit!

I think if we drill down to the real issue here, it's really this: How much control should an instance have over what it's users see and interact with?

Some folks want to be the ones to decide for themselves what they see and interact with, with minimal to no interference by the admin.

Other folks don't want to see or deal with objectionable material, and want the admin to ensure they don't have to.

I think the answer is somewhere between, myself. As others have pointed out, botnet instances and instances with illegal material should be defederated without mercy. Instances with little to no moderation and users that are causing problems on other instances should be considered for defederation too.

Whether to consider other cases of defederation is where it gets greyer to me. An instance that has communities that post questionable stuff, but doesn't break that instances rules, is one such grey area. And to be clear, I'm making the case that those communities and questionable stuff live on that instance.

To other folks, that wouldn't be grey, and I guess that's what we're really discussing.

I don't understand the point you are trying to make. On the one hand, you say:

Defederating entire instances does not stop bad actors, but an active strongwilled community does.

This makes me think you are saying not to defederate because it would be better to call out bad behavior - interact with the bad actors and point out their falsehoods, hate, etc. But on the other hand, you say:

I don’t interact with them. I don’t provide them with any value.

and

It’s not our responsibility to moderate other instances.

These make me think you are saying just ignore them. And if we're going to just ignore them, how is that different from the perspective of the bad actors, from defederating? How does not moderating and not interacting stop bad actors?

This is all new to me, I don't know the best use of defederating, but I didn't follow the argument you were making.

Defederation deprives me of the ability to respond to people from defederated instances, not only on their instance, but everywhere. Case in point, I'd like to respond to this doofus from lemmygrad, but I can't, because lemmygrad is defederated.

Agreed. While there are cases where de-federation is needed it's an option of last resort. The risk of balkanization is too great.

Yes it does. That’s why it’s in the spec. They built it for a reason.

Imagine 1,000 Russian bot and fascist trolls servers with 100,000 users each. There’s 20 major Lemmy servers now. And you expect us all to self-defend and block this onslaught?

Also: email servers work this way now. They “defederate” known spam servers all the time. For the same reason.

I think I agree. I'm uncomfortable with the idea of going too far, too fast in defederating from servers which have people whose views some users here find objectionable. There are some cases where it's fairly clear-cut, and I'd feel comfortable. Outright fascists and MAGA servers. Bot-farms. Ones designed to disseminate disinformation and fake news.

If their users are interacting with us in ways which are hostile, disruptive, etc. then we should defederate, obviously. Though that will only go so far given open registrations.

But this server is not, as far as I'm aware, intended to be of a specific political hue, nor primarily about politics. So, I think we should be cautious and careful, partly because I've seen 'tankie' used to describe such a wide range of views, often with very little in common between them.

I think from our instances view point it doesn't matter as much because as you said we have open registration. Having that open registration is what has led to us being defederated by other instances however with a longer approval process. So in the lens of the whole, they make use of that.

As for the purposes of us defederating an instance it would be

• bots/spams as others have said
• legal issues as part of the federation is importing that content posted by that instance (so I think having references to that content could make us vulnerable to issues with it). Again not sure on this one but for performance reasons could see the implementation being as such.

I agree with you.

Shit centrist take.

If a community doesn’t want to interact with a different one, let them not.

This whole “awww you shouldn’t ban people from the community, it’s just words so just ignore them” reeks of the shit Nazis complain about after getting banned.