Dunno man, I‘ve got the double red triangle exclamation mark for „very low reputation“. Certainly it „warns“ unsuspecting people and I had already at least one commenter, that mistrusted my commentary just for that. Cool. Cool, cool.
Don‘t have an account — do I also have an attitude, it doesn‘t show any publically?
Either way, branding users like that, intransparently, by machine logic, I certainly am not convinced.
There’s a bunch that goes into the reputation score, it’s a combination of upvotes/downvotes as well as other factors. PieFed also has a stronger slur filter, where instead of replacing with removed, it outright doesn’t accept the comment if you type the phrase.
All in all PieFed is built specifically to cultivate echo chambers in ways that go beyond what Lemmy limits itself to, by design.
It’s a mod tool. Hexbear and Lemmygrad are default blocked and instance admins have to manually unblock them if they want to federate. Further, if someone blocks a user, lemmy users can reply to the account but it won’t show on Piefed for anyone, not just the user.
I never said they were blocked by default, though. You’re taking issue with Lemmy.ml as a test instance despite Lemmy lacking in the tools for censorship PieFed has.
It does give a warning system for heavily downvoted people, the majority of whom are genuinely just trolls. I can’t tell you how useful it is to isolate trolls and spammers at an admin level.
Well at very high downvote rates, you simply can’t downvote anymore. It’s a mitigation against downvote trolling. There is also a hidden reputation score for instance admins which logs how much you also get downvoted.
Why am I forced to have certain voting patterns in order to be accepted as a trustworthy community member? That‘s arbitrarily authoritarian in order to subdue deviant behaviour without good reason or explanation before engaging in said community, which ux design displays up- and downvoting in the same manner? I‘m not being warned of such consequences for this arbitrary branding rule. Is excessive upvoting branded, as well? If so, in which way? If not, why not? It‘s dumb, that‘s what I‘m getting at.
I do this regularly. The „unfortunately“ part is not something I am willing to take ownership of and see it as a weakness of the platform. I also tend to be controversial in .world communities and I detest software that‘s trying to force me to behave in certain ways that are neither transparent, nor neutral in their judgment of my commentary.
That’s the same for up- and downvote balance. it’s bullshit. be it attitude or reputation.
I do this regularly. The „unfortunately“ part is not something I am willing to take ownership of and see it as a weakness of the platform.
I meant “unfortunately” in the specific context of the what’s happening here. I accept that it is somewhat of a problem here, but 9 times out of 10 - people who have been downvoted heavily tend to be trolls and spammers.
The mass downvoter gets their ability to downvote blocked.
If downvotes are such an issue then disabling them would be better right?
Downvoting isn’t considered inherently anti-social. It ensures that people use downvotes (or are more likely to do so) with the spirit of the system in-mind - off-topic content, trolling, spamming, etc. That said, I would simply disattach it and make it implementable at the instance level myself personally.
True. I suppose you could get to that reputation. Maybe it also accounts for time between each downvote. So I can imagine someone who doesn’t upvote but occasionally downvotes, and downvotes in a fair way. I don’t think they should get penalised. Perhaps it already accounts for this as someone like that is obviously very different than a spam-downvoter.
Once an account has made a few votes, an “attitude” is calculated each time they vote which is the percentage of up votes vs. down votes.
People who downvote more than upvote tend to be the ones who get in fights a lot and say snarky, inflammatory and negative things. If you were at a dinner party, would you want them around? By reviewing the list of people with bad attitudes you can make decisions about who you want to be involved in our communities.
All these accounts have been downvoting a lot (Attitude column) and receiving some downvotes (Rep column). Their profiles are worth a look and then making a decision about whether they’re bringing down the vibe or not.
Are you taking about the attitude where I’m currently sitting at 92% even though I don’t ever say the popular thing?
I don’t even know how it works but why is mine there?
Dunno man, I‘ve got the double red triangle exclamation mark for „very low reputation“. Certainly it „warns“ unsuspecting people and I had already at least one commenter, that mistrusted my commentary just for that. Cool. Cool, cool.
Don‘t have an account — do I also have an attitude, it doesn‘t show any publically?
Either way, branding users like that, intransparently, by machine logic, I certainly am not convinced.
There’s a bunch that goes into the reputation score, it’s a combination of upvotes/downvotes as well as other factors. PieFed also has a stronger slur filter, where instead of replacing with removed, it outright doesn’t accept the comment if you type the phrase.
All in all PieFed is built specifically to cultivate echo chambers in ways that go beyond what Lemmy limits itself to, by design.
I’ve never have a comment that hasn’t been posted so can you show me some code to prove what you’re saying?
Here’s a snippet from the code regarding blocking comments if a phrase is set to blocked. Credit to @edie@lemmy.encryptionin.space for finding this.
Are any phrases set to blocked by default?
No, not that I know of. It’s empty by default.
It’s a mod tool. Hexbear and Lemmygrad are default blocked and instance admins have to manually unblock them if they want to federate. Further, if someone blocks a user, lemmy users can reply to the account but it won’t show on Piefed for anyone, not just the user.
But there’s no phrases that are blocked by default because that’s what we were talking about.
I never said they were blocked by default, though. You’re taking issue with Lemmy.ml as a test instance despite Lemmy lacking in the tools for censorship PieFed has.
I’m taking offense to my donations to Lemmy being used to run this echo chamber.
Attitude is just how much you upvote vs. how much you downvote.
But why?
So how’s that a social credit score?
It’s the one that auto hides or warns others on piefed just like reddit is doing.
It does give a warning system for heavily downvoted people, the majority of whom are genuinely just trolls. I can’t tell you how useful it is to isolate trolls and spammers at an admin level.
Which is a majority of the reasons people get banned or comments removed from instances but is still called by them censorship.
However that process is a form a censorship as people can and will game the system like on reddit.
Just like how the reply function gets disabled without any mention to user if are blocked
People get removed from instances for being heavily downvoted on lemmy? Lemmy doesn’t have a base way for admins to isolate that, to my knowledge?
I’m sure Rimu could add in a notification for that.
Why would they when it’s the entire point of the system?
On some instances and comms absolutely
https://lemmy.world/post/41164530
But mostly was meaning people go to troll ml then call censorship when action taken
No, that was someone banned for downvoting - not for being downvoted. Difference.
Yeah, I know people can get banned for downvoting in some communities and instances.
Well at very high downvote rates, you simply can’t downvote anymore. It’s a mitigation against downvote trolling. There is also a hidden reputation score for instance admins which logs how much you also get downvoted.
Why am I forced to have certain voting patterns in order to be accepted as a trustworthy community member? That‘s arbitrarily authoritarian in order to subdue deviant behaviour without good reason or explanation before engaging in said community, which ux design displays up- and downvoting in the same manner? I‘m not being warned of such consequences for this arbitrary branding rule. Is excessive upvoting branded, as well? If so, in which way? If not, why not? It‘s dumb, that‘s what I‘m getting at.
You’re talking about attitude here, not reputation. Two different things.
Unfortunately you went into a partisan community a few months ago to argue with them and got heavily downvoted for it. That’s reputation.
I do this regularly. The „unfortunately“ part is not something I am willing to take ownership of and see it as a weakness of the platform. I also tend to be controversial in .world communities and I detest software that‘s trying to force me to behave in certain ways that are neither transparent, nor neutral in their judgment of my commentary.
That’s the same for up- and downvote balance. it’s bullshit. be it attitude or reputation.
I meant “unfortunately” in the specific context of the what’s happening here. I accept that it is somewhat of a problem here, but 9 times out of 10 - people who have been downvoted heavily tend to be trolls and spammers.
Wait the user downvoting gets blocked or the user constantly downvoted? If downvotes are such an issue then disabling them would be better right?
The mass downvoter gets their ability to downvote blocked.
Downvoting isn’t considered inherently anti-social. It ensures that people use downvotes (or are more likely to do so) with the spirit of the system in-mind - off-topic content, trolling, spamming, etc. That said, I would simply disattach it and make it implementable at the instance level myself personally.
But the system is specifically designed to punish those who downvote if they don’t upvote enough
True. I suppose you could get to that reputation. Maybe it also accounts for time between each downvote. So I can imagine someone who doesn’t upvote but occasionally downvotes, and downvotes in a fair way. I don’t think they should get penalised. Perhaps it already accounts for this as someone like that is obviously very different than a spam-downvoter.
Would have to go through the code but most likely just ratio of up to down
Nvm didn’t need to as it’s part of their join description
https://join.piefed.social/2024/06/22/piefed-features-for-growing-healthy-communities/
Find people who downvote too much
Once an account has made a few votes, an “attitude” is calculated each time they vote which is the percentage of up votes vs. down votes.
People who downvote more than upvote tend to be the ones who get in fights a lot and say snarky, inflammatory and negative things. If you were at a dinner party, would you want them around? By reviewing the list of people with bad attitudes you can make decisions about who you want to be involved in our communities.
All these accounts have been downvoting a lot (Attitude column) and receiving some downvotes (Rep column). Their profiles are worth a look and then making a decision about whether they’re bringing down the vibe or not.