This is an important thing to understand on Mastodon: you don't see all the replies to a post.

It's possible for evil people to propagate terrible hatred that only shows up for the person being attacked, but isn't visible to you.

See explanation from @mekkaokereke here: https://hachyderm.io/@mekkaokereke/111012743709881062

When Black users tell you there's racism, it doesn't matter that you didn't see it happen.

#Mastodon #MastoTips #Fediverse #Racism #Sexism #Misogyny #Homophobia #Transphobia #Hatred #Safety #Allyship

mekka okereke :verified: (@mekkaokereke@hachyderm.io)

@zachnfine@mastodon.social @JamesWidman@mastodon.social @Sablebadger@dice.camp @staidwinnow@mastodon.social @Jorsh@beige.party From your screenshot above, and the scenario I listed below, imagine if: 1) the racist user is on "Nazis dot social." They have 5000 followers. 2) the Black user is on "Good people dot social" 3) the nazi replies "Followers only." The scenario: https://hachyderm.io/@mekkaokereke/111010421955145872 None of the good people would see the gore images, or the subsequent pile-on.

Hachyderm.io

2/ This is a feature that needs improving, I think.

Three ideas come to mind (cc @Gargron )

1. Mark every post with global # of replies to remind everyone they're seeing a subset (I don't think this is a big help though)

2. Allow users to disengage from threads, exit a conversation they don't consent to, not just block users/instances. Not sure how this would work tho.

3. Make "followers only" and tagging a non-follower mutually-exclusive.

Hmmm.

3/ Well over 700+ “retweets” (I know, I know) but just received my *first* toxic reply. Nice change from other sites.

(Blocked and reported and already forgotten.)

@c_9 @Gargron you should be able to block "followers only replies". Like you do not accept someone to reply to your post as follower only

@c_9 @Gargron it sounds like
1. Person A posts
2. Person B replies
3. Troll C, who is already blocked by A, replies to B with harassment. A can't see it, B wonders why A doesn't care

It seems like the problem is that blocking is all on the receiving end, and we let folks reply to those who block them. I understand technically it would be hard to enforce preventing replies to blockers, but we could also notify B in this scenario "A has blocked C. Would you like to as well?"

@c_9 @Gargron still requires leaking your block list though 🤔

@TimPurdum @c_9 @Gargron wait ... blocking user A doesn't mean A can't see my posts?
If so, what's the difference between Block and a permanent Mute?

Edit: apparently, users we block DON'T see our posts. Thx to the commenter who tested it. 👍

@deborahh @TimPurdum @c_9 @Gargron I'm pretty sure that if you block someone they can't see your posts?
@deborahh @TimPurdum @c_9 @Gargron I'm pretty sure that if you block someone they can't see your posts?

@anarchopunk_girl @TimPurdum @c_9 I thought I read that it was different on Mastodon than on xitter?
/cc @Gargron

Edit: people we block don't see our posts, according to a commenter on this thread who tested it. Thanks 🙏

@deborahh @TimPurdum @c_9 @Gargron I'm pretty sure that if you block someone they can't see your posts?
@TimPurdum No, the OP was about C doing tagged follower-only posts, so only A and follower of C are seeing it.

@c_9 @Gargron

1. Is a client thing.
2. What would be the result of this operation? Maybe use https://github.com/mastodon/mastodon/issues, if you are using Mastodon.
3. Breaks the symmetry of the levels, or you remove tagged only as well. And I'm pretty sure this is a source instance operation.

@c_9 @Gargron it’s astonishing that #3 isn’t a thing already.

@c_9

Agreed. Perhaps also a "This thread contains replies from users you've blocked/muted" notification? Or a more general "There are replies you aren't seeing" notice, possibly clickable for details.

And I think, for this and several other reasons, when I view a post, my instance should go fetch *all* the replies to it. And, hopefully, get a notice when replies exist that my instance can't see because they're from a defederated instance.

@c_9 @mekkaokereke

The tragic byproduct of this is that just by virtue of having a prominent account you can be amplifying racism you can't even see. @jwz was exactly right when he said "how it should work is, I own the replies to my post". Which means, the system makes a concerted effort to show _me_, at least, all of them so that I can reasonably moderate - and be held accountable - for what I amplify.

@mhoye No, it can't. Because those posts are only visible to follower of the troll. So no amplification, only existing connections work.
@mhoye @c_9 @mekkaokereke @jwz
SFAIAA The account / server making the initial top-of-thread not-a-reply post will see all subsequent replies where they are included on that reply, yes?
@c_9 This is a fundamental problem. Can imagine if the KKK could turn invisible? Nightmare scenario.

@c_9 @mekkaokereke

(lightbulb) Thanks!

This explains a number of posts I've seen from people that I follow, where they're obviously distressed about reactions they've gotten and where I have no idea what they're talking about.

@paulvg But there should be a key giveaway: a tagged user you don't know and who have not been previously involved in the conversation. Plus that post suddenly is follower-only.

In the end you are always limited to "trust by default" or "not trust by default". Mastodon defaults to "trust by default", so you block stuff away.

@c_9 @mekkaokereke (YES AND this is generalizable: abuse is often purposefully invisible except to its targets, and one should keep this in mind when evaluating claims of abuse.)

@c_9 @mekkaokereke
Thanks for posting this. Yep, lots of imperfections in the fediverse that need work. The lack of algorithms driving what we see is, IMO, not an imperfection. That being said, your post is important. Might a hashtag or two be useful to expand reach? Oh, and I do enjoy the edit feature (hint 😉 ), even thought that is also open to abuse.

Again, thank you for the post (which I only saw because someone I follow, @OrionKidder, reposted). All best!

@c_9
thank you
That explains a lot.
@c_9 oh now I understand why I can see there are replies to something but I can’t actually see replies!

@c_9 @mekkaokereke

Huh, I wasn't aware of that. Explains why one of my favorites, Lexi Alexander left. I didn't see anything, but her replies were obviously very distressed. So she must have been seeing awful behavior.

I don't want to see awful behavior, but I don't deny people their own experiences either.

@c_9 @mekkaokereke

I wonder if that is a flaw of the system, it is certainly misuse…

@c_9 @mekkaokereke How about a screenshot of the offensive toot and the reply, calling out the racist. There is a need to educate on the many shapes that racism can take. Just a suggestion. #EnoughIsEnough
@c_9 @mekkaokereke In reading this post I see functions I have never seen before, and I see now how that can happen, but a report from anyone could also get that whole instance banned. This is not to take away from someone's pain in experiencing this, but if they report this user and thread to their own instance then their instance can take action and if the host instance of the harasser takes no action then that whole instance could be banned. Maybe I'm naive.

@c_9 @mekkaokereke sorry to jump in, but it seems like this is the flipside of mastodon promising something it can't really deliver, i.e. privacy.

we've collectively been (kind of) lured by the prospect of “confidential” posts & replies which are inherently not (anybody who has played with activitypub knows this for a fact).

basically, “confidentiality” settings gives us the worst of both worlds: a false sense of privacy & (as we privileged folk collectively discover) a vector for abuse.

@c_9 @mekkaokereke one way to fix this might be to ditch any pretense of confidentiality & instead present these settings as “prefferred audience”. meaning these are the people whose timeline you're aiming for, but there's no guarantee beyond that & anybody who clicks the original post can read all the replies because nothing can be private on this network, period.

at any rate, we need to work harder to stop harassment on the fediverse.

@c_9 @mekkaokereke
🥥 Eye, for one, did NOT know about that. 🥥
@c_9 @mekkaokereke that's why fediblock and other defederation initiatives are very inportant
@c_9 @mekkaokereke This is appalling and all too easy to imagine. But aren't all federated servers obligated to honor the basic "no Nazis" requirement? It's hard to understand how a server hosting racist abuse can remain connected. (Sorry, I missed the larger thread. Reading now.)

@jodystillwrites @c_9

"Obligated to" and "actually do" are often different things, especially when it's hard/complicated/expensive to do the right thing.

Suppose a new Nazi server pops up tomorrow. Let's call it "anti-woke dot social" they immediately start harassing users on the biggest instances, and promptly get defederated. So far so good.

But... How would a tiny instance even know that this happened? How much work would admins have to do to stay up to date?

It needs to be easier.

@jodystillwrites @c_9

Most admins don't want an extra, (unpaid!) part time job, doing the same nazi investigations that other people are already doing.

Better to just check a checkbox.👍🏿

☑️I would like to use the auto-updating nazi filter.
⬜ I will do all my own nazi hunting. I am an expert.

(Should be a radio button, but hey).

@mekkaokereke @c_9 You're right...it only takes minutes for a pile-on to occur and traumatize its victim. I hope a real-time way to stop this threat evolves soon.
@c_9 @mekkaokereke
I did not knew that as well, hope that issue can be solved asap.
@c_9 @mekkaokereke

I was just thinking today that I see far more misandrist messages than misogynist ones, this above explains why.
@c_9 @mekkaokereke to be quite honest, I do not think this is wholly a technical issue. White people will say this even when the racism is right out there for them to see. There are white guys out there right now saying they've never seen any racism/sexism/etc. in open source.
@nev @mekkaokereke 100% agreed. Modifications to features and moderation are (I think) necessary but not sufficient, and even an imaginary perfect app/fediverse will have mostly-white mostly-guys believing their privilege over lived experience.