Idea: :dont_at_me: as a permissions level for a post.
Can still be boosted and favorited, but servers should prevent replying, and a server implementing this will reject any incoming replies.
@Elizafox I mean, that's certainly simpler, although I can see a couple situations where that's insufficient.
(The biggest situations I can think of are when it's instance moderation that does a :dont_at_me:, and then replies are happening that may need to be dealt with by said moderators.)
@Christian I mean, the entire permissions model currently in use is messed up, and Pleroma's handling of replies just shows that even more (it puts any reply that anyone you follow makes to anyone in your timeline, even when they're in reply to something that you can't see).
@bhtooefr Alternatively close thread.
But these suggestions will make the system very heavy; it will require one lookup for every incoming message that is in-reply-to something and that might be a bit much for high-traffic instances.
@SuperFloppies Mastodon's already doing many lookups for every incoming message, for timeline insertion, FWIW
@bhtooefr Well, that is a design issue and not strictly needed for the implementation. But in that case, if the new lookup fits into the old ones it adds just a branch to the code. I wouldn’t assume that tho.
Replying to a very old post will have the highest cost. “One lookup” sounds small, but remember that no implication is made on the amount of I/O required for that one lookup.
Cache effect does not exist for old data.
@bhtooefr “design issue” meant to say “design detail”. Not sure how issue came out there.
@bhtooefr That's an interesting idea. I do wonder however whether this could be abused. For instance, if someone posts something and then I can post something with "don't @ me" as a first comment (which will be shown to anyone that clicks on the original toot) with some factually wrong information, mean stuff, etc.
- then nobody can reply to it and set things right...
And this is just one of many examples that I can think of where abuse could be enabled by this.
@natanji Note that the "don't at me" flag would only apply to the post that it's set to apply to - so your post would be a subthread that couldn't be replied to, but the original post could be replied to (unless you were a moderator or the OP, and explicitly aimed the "don't at me" flag at the root post of the thread).
But still, interesting point.
@bhtooefr I guess I wonder if it's a good idea in the first place to try to treat toots like posts in a forum where you are a moderator. Or like personal blog posts where you can turn the comments off. I get that this seems to at first just give more control, but in an ecosystem of a social network - which is what Mastodon actually is - I see lots of problems to give such asymmetrical control tools to users.
@bhtooefr Nobody would allow users in a forum to just make new immediately locked threads where they post their views of things without any possibility to reply.
The "don't @ me" is a social request that doesn't require a forcible technical solution (with bad side effects like what I described), but a social one.
If someone posts bad nazi shit, I wanna be able to counter them publicly; not have a technical system in place that prevents any kind of dissent.
@Satsuma Yeah, so?
@bhtooefr Don't at me is one of the most frustrating notions that I came across on twitter.
If you just want a one-way channel for broadcasting, why not post your thoughts to a platform designed around that (e.g. your personal blog/website without comments) instead of on a platform designed for encouraging discourse?
It's like logging into an IRC channel, posting something controversial and then getting mad that it sparked a discussion.
Server run by the main developers of the project It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!