I still think Algorithms™ would be good for fediverse birdsite clones if they were based on a plugin architecture so users could choose & configure exactly what's done to their timeline.

Example ideas:

· keyword filter for specified posters

· "delay buffer" for those who post/boost 900 times in an hour then go silent for 3 days

· only display posts above a given engagement threshold (e.g. for chronic shitposters with a few really good ones)

Think of it as guitar pedals for your timeline.

@deutrino totally agree, but user controlled or instance wide settings?

@santiago user controlled! :)

@deutrino these could all be done with MRF, but per-user configurability still isn't part of that unfortunately

@velartrill I'm curious if you know whether MRF could be used to force outbound subject/CW, and/or to add one to boosts. I haven't dug into MRF at all yet.

@deutrino i only know of it being used as an inbound filter. what you describe sounds like extremely user-hostile behavior tho so i wound hope not

can rts even have separate subjects from the tweet they're retweeting?

@velartrill I don't actually know whether adding a subj to a RT is technically possible, no. I was mostly musing on whether this sort of thing was possible for moderation of e.g. habitually crass users (on servers where they aren't just banned outright), or for edge cases like porn instances where mandating outbound CW might be appropriate or even desired (since a lot of ppl block the ones that don't enforce a CW/subj line on noodz)

@deutrino as far as i know you can't have multiple subjects so this would mean taking away the user's ability to set them himself, completely defeating the purpose of the field. anyway, i don't think these are particularly valid use cases. if a user won't follow the instance's rules, you ban them.

@deutrino once you've established bad faith and willful noncompliance, there's really no reason to fuck around. being wishy-washy and irresolute dealing with people who behave like that just signals weakness to them and encourages them to keep causing trouble

@velartrill not everyone is going to run their instance like you though. in fact I think diversity in moderation style may even be an asset to the wider fediverse.

@deutrino gonna be real here, failing to understand the distinction between good- and bad-faith behavior isn't a "moderation style," it's just lack of basic leadership skills. people who act in bad faith aren't going to just suddenly turn around and become model netizens if you coddle them enough; it's why free speech zones are such unmitigated hellholes.

i mean, if you want to enable people who want to abuse your resources, go ahead, no one's stopping you, but don't expect others to implement features that would encourage that kind of mismanagement and help you create an unpleasant and unrewarding environment for your good-faith users. (well, except maybe gargron, he seems to love doing shit like that.) refusing to exclude people who violate your own rules is one of the faster ways to wind up blocked by everyone but the free speech zones tho

@velartrill @deutrino

yes, MRF can rewrite outgoing posts too.

after all, they're incoming when the local user submits them and they are translated to activities.

@kaniini @deutrino ah, good to know, thanks!

@kaniini @velartrill holy shit a kaniini sighting, hope you're well

@deutrino @velartrill i'm fine, just in the process of moving to holland, have been busy with that.

@velartrill you can skip the lecture, i've been in more than one rodeo over moderation style in my last *counts* >25 years on what the kids call "social media" these days.

if someone running an instance like SPC or FSE has ethical problems with banning users who habitually step on taboo lines (Truly Malicious or not, which I guarantee would be debated), a mandatory outbound CW may be useful.

similarly a global CW for somebody running e.g. a porn node who wants to be a slightly better neighbor.

@velartrill personally, I'd just ban the bad actors, but I was pondering how the spicier instances might be able to coexist a bit more peacefully after some chud invaded a thread I started. heh

@deutrino ultimately it comes down to enforcing policies about how their users interact with users of other instances, or taking moderation action against individual users of other instances. also i think you'll find that the free speech zones are even more opposed to compelling speech than censoring it

@deutrino @velartrill

tired: filtering as mitigation

wired: consent baked into the protocol from day 1, so that participation is defined to the set of people you want to actually talk to

due to Gargron having to stomp on the gas without thinking about things, we know which way ActivityPub went.

@kaniini @deutrino something i've wanted for a long time is the ability to like, write graph expressions that control who in my social graph can see, follow, or interact with me. e.g. "allow friends^1 of ppl in group Cool to follow me excluding friends^3 of ppl in group Uncool." that kind of fine-grained control would have been really cool, and would let people scale all the way from a public twitter-style experience to a private facebook-style experience all on the same platform.

maybe someday we'll see a protocol that lets you have that kind of control

@kaniini @deutrino oh cool! i also moved to europe this year, high five

@velartrill @kaniini @deutrino this is what Google tried to do with their "circles" thing in G+ IIRC

@deutrino A filter to bump posts from low-activity users to the top of my TL has been on my Fedi wishlist since day 1.

@videomurder @deutrino @kaniini yeah i really liked circles, it wasn't abstract enough tho