Mastodon: Enable Twitter-style Reply Controls on a Per-Toot Basis

Created on 8 Sep 2020  ·  10Comments  ·  Source: tootsuite/mastodon

Pitch

Twitter's reply model has been extended with some LJ-like features.

Replies to a tweet can now be restricted to:

  • Replies only from accounts \@-mentioned in the tweet
  • Replies only from accounts followed by the sender of the tweet and those \@-mentioned in the tweet
  • No restriction

Something similar to this was proposed in #8565, from two years ago, but the Twitter implementation is more robust.

One of the objections to this was the fear that a user could \@-mention a user, and disable replies. Twitter's implementation is aware of this and allows any account \@-mentioned to reply to any tweet.

Motivation

This allows no-comment toots by setting the \@-mention only reply setting, and not mentioning any other accounts.

Communities, especially marginalized communities, need a way to have discoverable conversations, but limit posting access to members of the community (via following.)

This can be used to assist in moderation so that, for example people involved in the conversation don't have to spend time explaining background (ie 101, google the topic, etc) or passive aggressive "reply-guys".

This expands and empowers users so they don't have to use brute-force blocks or mutes.

Most helpful comment

The purpose of the feature is to let people control who responds to a post, such that a person does not have to make a post followers-only to avoid abusive responses.

This is the equivalent to a blog post without comments.

If a reader feels compelled to respond to a blog post without a comment facility, or toot with replies off, they may do it from their own post on their own blog, or social media account.

If they disagree with the content, they can mute or block the account. If the content violates that instance's ToS, it can be reported.

I'm still not clear on your objection.

All 10 comments

Heartily agreed. This has the potential to eliminate "reply-guy"-ism in one fell swoop.

In addition, one of Mastodon's selling points is that it has more privacy- and moderation-friendly features than Twitter. This is an actually useful feature Twitter has introduced which new adopters will find missing if they move to Mastodon. Mr. President, we must not allow a mineshaft gap!

It's a valuable feature (although I'm slightly afraid this could worsen “echo chambers”), and something we are investigating. However, this requires major changes to the protocol, and there are a few caveats to consider:

  • implementations/server which do not get updated to support whatever protocol changes we come up with (or maliciously ignore them) will still be able to reply to any toot. The best we can ever do is to have those replies not appear on “well-behaved” servers implementing the protocol
  • this dramatically changes how threads can be resolved: so far, toots can be processed before processing their ancestry, but if we need to check whether a reply is allowed, we probably need to fetch the whole thread before saving things to database, which means we probably will have to limit the length of discoverable threads or something
  • the list of people who are blocked or follow someone isn't necessarily public, so the originating server would have to vet each toot independently. This means the originating server can apply different rules than those advertised or understood by other servers, which may lead to replies being dropped for no obvious reasons. One could make an exception that mentioned users do not need the originating server to vet the reply, I guess.

Also, there's something that I'm not sure about: what is Twitter's behavior when replying to a reply? Is the reply policy locked to that of the original post, or can it be changed down the road?

EDIT: also, note that we are investigating other ways to handle replies, but it's going to be a long road, and the caveats above do remain

Thank you for your explanation of the technical issues involved!

In my opinion, preventing harassment is more important than preventing "echo chambers", which users can choose to avoid by simply…following a wider range of people.

In my opinion, preventing harassment

No offence, but isn't that the purpose of the Block facility?

No, because blocking happens after abuse. Limiting who can reply reduces the need to block.

Exactly

Let's keep +1 comments to a minimum.

Personally, I think this introduces a brilliant vector for offensive posts. If this were to get implemented, I could say with confidence that someone is going to post some racist propaganda, then prevent people from responding to it.

I'm mixed. This may be good for fixing your problem, but it's definitely going to lay the foundation for a lot more problems, which is being grossly underestimated.

Following this, it may not be appropriate to add features that limit the social aspect of a social platform.

This may be more trouble than it's worth. I'm not sure.

The purpose of the feature is to let people control who responds to a post, such that a person does not have to make a post followers-only to avoid abusive responses.

This is the equivalent to a blog post without comments.

If a reader feels compelled to respond to a blog post without a comment facility, or toot with replies off, they may do it from their own post on their own blog, or social media account.

If they disagree with the content, they can mute or block the account. If the content violates that instance's ToS, it can be reported.

I'm still not clear on your objection.

If this were to get implemented, I could say with confidence that someone is going to post some racist propaganda, then prevent people from responding to it.

Do we have any evidence that people responding to abusive tweets/toots with criticism makes the person less likely to post abusive tweets/toots in future?

This may be good for fixing your problem, but it's definitely going to lay the foundation for a lot more problems, which is being grossly underestimated.

Do we have any evidence or experience of this from Twitter, where they have this feature? Is it being abused by racists/other nasty types to prevent people from responding?

My instinct says that people posting abusive stuff will leave replies on because they want people to reply, to get them more attention and outrage.

If someone is posting something abusive, they should be reported. I don't believe that allowing people to turn off replies will protect abusive posters.

Personally, I think this introduces a brilliant vector for offensive posts. If this were to get implemented, I could say with confidence that someone is going to post some racist propaganda, then prevent people from responding to it.

i would suppose if i find an offensive post on my TL, that i block them and report them instead of engaging with them. the point from what i gather from this enhancement request is that such person is preventing anyone from responding to that specific offensive toot. sounds good to me, i just block and report. i dont want to engage with that person anyways but also prevents anyone else from engaging them unless specifically mentioned.

  1. "observers" wouldnt get trigger happy to respond to offender, they can shout, no one would respond to them until reported and probably banned.
  2. brigading would be reduced as now people would not have a way to organize against someone except by way of reporting them.

just my 2 cents

Was this page helpful?
0 / 5 - 0 ratings