After a couple days of discussions about the newly implemented vote quota, I’m kinda exhausted. It seems like a situation we won’t ever agree on. Me personally, I don’t want to argue like this over a piece of software that I have high regards for. It tears us apart, where we should work together.
It’s okay if there is a quota on piefed.social the instance.
It’s not okay if there is a default quota of 240 on PieFed the software - and thus for all instances.
I suggest it should be implemented like this:
- It should not be a default value
- It should be an empty input in the admin interface, where instance admins can set a vote quota if they want to, or leave it empty to disable the vote quota.
- The /about page should display the set vote quota.
That way all instances can decide for themselves and users can see the instances’ vote quota transparently.
You must log in or # to comment.
This isn’t the first time piefed has done some very stove-piped feature development. The same thing happened with voting agents, which were quietly removed after a bunch of backroom conversation and forum politics. It’s a bad look all around.
If the dev wants to implement this feature as is, and won’t remove it from being opt-out, then I will just leave Piefed and go back to Lemmy. I hope Ada over at Blahaj knows about this so they can remove the limit from their instance.
The trouble I have with Lemmy.ml (this is relevant, I promise you:-) is not how it bans all criticism of Russia, North Korea, and China, but in its lack of transparency to never ACTUALLY STATE that this is their “real” set of rules, which differs rather profoundly from its own provided statement of being “A community of privacy and FOSS enthusiasts, run by Lemmy’s developer”. Overall that reads as extremely disingenuous to me - while to them, OF COURSE they are doing the “correct” thing, in ALL matters, and yet… while I am not asking them to provide a justification, it would be nice to acknowledge in the light of day what policies are actually in place? It scares people away to have to discover that the hard way, and many seen to not land on other Lemmy instances afterwards but rather go back to Reddit or Bluesky instead.
Thus likewise I am against how this voting scheme is not currently implemented in a transparent manner. e.g. this noticeably reduces the Newbie-friendliness of an instance, does it not? Users coming over from Reddit will not expect it and, already being overwhelmed by “there being too much to read” (I don’t personally agree but that is what THEY say, hence objectively true that this is their primary response), will get surprised and dismayed.
Also, if it helps to share, at first I wondered: what if the default were to be set to be a value higher than a human being is likely to reach (even someone recovering from an illness so indulging in social media more often than a usual balanced individual may casually & routinely do?). As a defense against bot accounts it may serve some value? Although then again, would merely limiting the bot be the best course of action here? Possibly detection and banning could be a better course of action, or even redirection into a honeypot, rather than steering their behavior to fall more into lines of human norms? So no, the more I think about it, the more that I see this not as a measure to separate bots from humans but rather to enforce conformity among human populations: neurodivergent or otherwise “unusual” people are told to get their own instance from now on, while this level of activity is all that is allowed in the main spaces (and even then, the votes cast and displayed internally within the neurodivergent spaces will be blocked upon entry into the main ones, having been throttled).
Anyway, the more transparently this can be handled the better.
I throw out upvotes like a drunken sailor. At minimum, if someone contributes positively (or at least on-topic) to one of my posts/comments, they get an upvote. It works as a “mark as read” on my end and makes the person feel seen on theirs.
I’m not sure I’d go through 240 a day, but I also wouldn’t want to have the added stress of rationing them out.
The whole quota system feels heavy-handed to me, but I’m not on Piefed so I got no horse in that race.
What if I just ask nicely for an upvote?
Ask and ye shall receive.
Can I have another?
Sadly, users of PieFed instances that implement this default without actively opting out will not receive your upvotes:-(. But thank you for your service!! You are kind to have offered them regardless!!
It applies to received votes too? I thought it was just a quota you can give (I haven’t been following this feature or the drama surrounding it, so forgive my ignorance as I try to get up to speed).
It looks like it yeah: https://codeberg.org/rimu/pyfedi/src/commit/7e10e92de2cf271088b76a31d725cad67afe08aa/app/activitypub/routes.py#L2335 - all votes, even those cast from Lemmy, Mbin, nodeBB, Mastodon, etc. would be filtered, it would seem? But don’t take this as authoritative on my part: I have not studied the code personally.
If I’m reading that right (my Python is rusty and this is the first I’ve seen of the PieCode), then yeah, it seems only
VOTE_QUOTAamount of my votes would be received by Piefed users.Pardon my French, but that’s horseshit lol.
Imagine if you were on an instance that allowed 500 daily votes, but then PieFed.social only federated in the first 240 of them. Along with every single other instance that uses PieFed software and does not override the default settings. Maybe some instances will choose to allow 1000? Others 100? Still others unlimited? This is like defederation: it complicates the entire process of sending and receiving messages using the ActivityPub Protocol to implement the Threadiverse ideal.
And PieFed.social with its 240 daily vote quota is no tiny instance!! It matters to the Threadiverse overall what THIS instance in particular does!!!
Worse though, none of the instances are going to say what their particular settings are. New users would be impacted, except they cannot see what those settings are so cannot make decisions based upon any reliable source of facts. This will lead to fewer incoming users overall, at a time when the Threadiverse is already in decline.
There was a good idea buried somewhere in this implementation, but it was extremely poorly executed. What should have been an opt-in feature, fully implemented before being put on prod, and transparently explained, was instead haphazardly done, and the Fediverse overall will feel its impact for years to come.
I read through some other links you dropped in another post and, yeah, there’s a good idea in there somewhere but it seems like it would be better handled case-by-case if vote manipulation is suspected.
I’m also not really buying the “These accounts control what you see” argument, especially given how many people simply browse by “new” where the score doesn’t matter for ranking at all. I don’t say that to suggest an ulterior motive, but simply that the problem seems blown way out of proportion and the solution feels poorly thought out and hamfisted.
Wouldn’t the outside user’s vote only be filtered on piefed.social if the outside user had also voted more than 240 times that day?
Or are you suggesting that a piefed.social user receiving a vote also contributes to that piefed user’s limit?
The former, one presumes, according to the code.
Then, based on the data, piefed.social would only potentially filter out the votes of a couple hundred people, which wouldn’t be noticeable.
It’s okay if there is a quota on piefed.social the instance.
It’s not okay if there is a default quota of 240 on PieFed the software - and thus for all instances.I suggest it should be implemented like this:
- It should not be a default value
- It should be an empty input in the admin interface, where instance admins can set a vote quota if they want to, or leave it empty to disable the vote quota.
- The /about page should display the set vote quota.
That way all instances can decide for themselves and users can see the instances’ vote quota transparently.
I agree
I’ve always believed, until this week, that if I posted something and it got upvotes, it was because people liked that I posted it. I’ve learned this week that because of some, it’s become a much less meaningful indicator. It’s been co-opted for unrelated purposes. I no longer trust it has any meaning. This is a letdown. Let’s add it then to the list of meaningless metrics like lines of code, story points, and StatCounter statistics.
If it remains an admin choice, I will keep it enabled for the two reasons that the behavior makes vote counts meaningless, which is unfair to the community, and that these empty votes add a ton of unnecessary traffic for every instance. I wonder what percentage of my server is used just to process meaningless votes. Each vote sends me a request to update the database. Each meaningless vote here requires me to send out a ton of requests to other servers. Why should I pay for this? Why should users see lower performance because of this behavior?
Please leave the setting, Rimu.
And I completely like the proposal.
I’ve learned this week that because of some, it’s become a much less meaningful indicator.
This is the part I don’t get. How does it become less meaningful? Because one user votes thousands of times a day, and several dozen users vote hundreds of times a day?
The fact remains that on each individual post or comment, any given account can only vote one time. If someone has multiple accounts and votes multiple times on the same content, that would be voter abuse and should be handled accordingly, but that doesn’t seem to be what this is about.
How is it unfair to anybody that different users vote at different rates? Everyone has the same options for every post or comment: upvote, downvote, or neither. (Side note: this is a ternary mechanism, not binary as I saw someone else describe it). If one person votes on a hundred different posts, their influence on any single post isn’t outsized compared to someone who votes on a dozen posts. Each one’s vote still only counts once per post or comment.
I think this has utility, especially in the comments, because it helps sort content loosely based on what members of that community want to see or don’t want to see. It’s subjective, sure, but it’s also organic, and honestly that’s how it should be.
Like others have mentioned, I wouldn’t want to ration my votes. I don’t think about how many I give out, I just upvote or I downvote or neither based on the content. No matter how much content I vote on, I still only have one vote per, so what’s the issue?
I can also see how this would have a chilling effect overall when users start rationing their votes. It becomes less of an infinite resource to dole out liberally and more something to spend wisely. We’ll have to start thinking “hmm, this is good/bad, but is it worth spending a vote on to cast my opinion?”
As a result, content won’t receive as many votes. And that’ll be disappointing, cause it can be gratifying to see that a lot of people agree with your comment, or that they think your post is worth more people seeing.
If you only ever get a handful of votes on anything, it might start to feel like “What’s the point? Why keep posting?” And then the chilling effect might not only effect votes, but even posts as well.
I’ve already seen at least one well-known user who posts great content say they might leave the threadiverse over this. Is that really what we want? Is it worth it just to prevent a handful of accounts from voting hundreds of times a day?
That post received only 16 upvotes, and many of those are likely to reflect the overall sentiment about the voting situation being disliked, or otherwise indication that the post is “relevant” to the community it is in.
I do not think of it as an indicator that thousands of people are running around the Threadiverse upvoting literally everything they see. In fact in Rimu’s various analyses he indicates that is definitely not happening. Maybe 2-10 people, but not thousands, hundreds, or even tens.
Actually, at that point the user may look indistinguishable from a bot, and thereby be worthy of a ban from your instance for that reason? It’s kinda an extremely low-key DDOS to do so!!
The VAST majority of people do seem to truly use voting as it was intended - to signal that more of such content should be continued to be added to the community?
So analogous then to the prom king & queen winning the votes. Ultimately it’s going to be the popular kids rather than people who truly contribute or impact their environment, but it does still have SOME kind of a meaning? (The trouble with this style of reasoning is that the likes of Peter Thiel, Donald Trump, Elon Musk, Hitler, Jeff Bezos, Stalin, Bill Gates, Putin, etc. are all “impactful” even if not “popular”. So impactful != good, necessarily.)
At the end of the day though, it’s your instance and you are entirely free to do as you wish with it. Again I think you may want to ban users that abuse the system in such a way - if they are making ALL votes meaningless then this is actively doing harm, right? (Also, someone could try reaching out to them and coaching them, but if they aren’t on your instance then it makes sense to leave that to their home server, and just protect your own users?)
Maybe threadiverse UIs should start having “mark as seen and hide”, so we don’t get “upvote to hide” behaviour?
Summit has “hide read posts”
Mark read on scroll and hide has been an option on Boost from the start. I wouldn’t consider a reader without it at this point.
“The fediverse is good therefore everything in the fediverse is good therefore everything must be upvoted” seems to be their reasoning.
Naah, that ain’t it.
Upvote my posts/comments if you want other people to also see the post, not because you want me to feel warm and fuzzy. If you want me to feel warm and fuzzy, comment with “Great post, thanks!” or “You are a fuckin genius rimu, always such insightful comments!”.
Voting controls who sees what content, it’s not kudos for content authors. If it was about the kudos we’d be designing the UX totally differently. Notifications when you get lots of upvotes on something, daily summaries of which if your comments got upvoted the most, who upvoted you the most, leaderboards of who got upvoted the most, etc etc. All of which sounds very toxic…
If you keep going in this direction, I’ll quit Piefed for good and return to Lemmy. You can’t force your ideas about how social media should be without considering its users. You don’t live on an island.
There needs to be a compromise. I agree with the others. Make the feature an opt-in for admins.
I disagree: votes are whatever the recipient makes of them.
Voting controls who sees what content
Sorting by New - even comments underneath a post - prevents the votes from controlling who sees what.
it’s not kudos for content authors
Again, that’s not up to you. Perhaps you meant that it “shouldn’t” be that way? But none of us get to decide how others actually feel about things.
More to the point, votes reflect an “engagement” - a post that people upvote is one that they are saying is relevant, whether they enjoyed the content or not, while a fantastic item of content that receives hardly any votes…
is content that will in the future go to other platforms.
I agree. I upvote just about anything that fits the community and I appreciate someone posting. I also post stuff explicitly hoping that other see it, and we have to face facts: many people just lurk and upvote, and there hasn’t been a good way to get around that.
Yeah i also like this vote quota. I would even go futher to completly remove it and find another system that can help us select content.
@jakylla@jlai.lu told me that tournesol are currently working on an fediverse algorithm where users categorize content.
I believe in something as flair.
If you want to make personal use of a different sorting algorithm, that’s entirely fine, but why take away the ability of the rest of us to use what we are used to?
Also, I think that sorting by All/New, or particularly New within the Topic Areas, is already a great way to discover content entirely independent of people’s voting habits.
I would completly remove voting system because it completly miss its point. It is not very useful.
It doesn’t help people finding content but tell people, company what we support on the largest userbase community : “I like this, i don’t like this. Most of users here enjoyed this post.”
“Yes, no, yes, yes, no.” I think this is a binary thinking.
Those reaction doesn’t help me at all. I think it is fine to enjoy or not the post and comment. But we can imagine something better than Reddit at shorting content.
For example, youtube’s video have also a like/dislike bar. Why are we following and using the same tool made by big tech ? IMO, this system is designed to shoot us with a small dose of dopamine that boost our ego and profil us for advertissements since we know what you like.
They were mostly created to held our attention on screen…and deprive us from our sleep.
Maybe, we can change that and try something else like https://tournesol.app/
So first of all, it would be nice if this hypothetical “better” algorithm already existed, before we start limiting votes here & now, on this instance? Don’t let perfect become the enemy of good?
Second, I think you’ve missed something wherein while each vote is individually binary, in aggregation they become a numerical value that is no longer binary - something that receives 1000 upvotes is significantly more popular than something that receives only 10.
Third, whether the current system is “good” or not is not something for us to judge? If people WANT something, then why prevent them from having it? Even if something else is “better”?
Sometimes people browse by New, and these people will not be impacted (on the side of receiving fewer votes than before this anti-feature was implemented), however many people desire instead to spend their time differently, and be alerted to what others have found to be “good”. This is obviously not a perfect metric for what YOU may enjoy, but… it’s a start?
We don’t need to engage in binary thinking ourselves: a system can be more than simply “good, so keep it as-is” vs. “bad, so now get rid of it entirely or at least start limiting it”, there are shades and nuances possible within all of this. Especially the knock-on effects, I mean events that transpire downstream and as a direct consequence of the upstream ones, even if logically they seem at first to be disconnected. e.g., what link is there between offering posts vs. votes - none? But if people leave an instance as a result of this change, then fewer post content can result, despite the seemingly disconnected nature of those two.
Voting makes the Threadiverse “welcoming”. You call it dopamine but however it is implemented, it makes people feel good to be here, and to contribute to a community. Perhaps even a wider, global community, beyond what they could reach if they merely went outside to touch grass irl?
So yeah, make that better system first, and then offer it to people as an opt-in feature. However, the point of this post was to discuss a mandatory new feature that removes existing capabilities, NOW not at some future date when your mentioned hypothetical opt-in feature has been deployed to a production environment where people can use it here on PieFed. So you might want to put your comment here over into the original post discussing this new voting feature? Because it does not seem terribly relevant to me for it to be HERE, discussing just how mandatory this new feature should be made as? … unless you truly do want the capabilities of others to be curtailed?
Why do you call it antifeature ? The vote quota is the equivalent of scaled feature.
Voting makes the Threadiverse “welcoming”. You call it dopamine but however it is implemented, it makes people feel good to be here, and to contribute to a community.
It won’t impact them. The vote quota is high and you miss the point. Ask yourself why it was created.
Let’s take another example. We organize a metting with several people and two people are completly monopolizying the meeting and sharing their idea.
What if we try to balance a little speech duration during the meeting ? What if we ask those users who speak the most to stop and let other sharing their idea ? That’s what we do in our anarchist meeting, we try to balance and limit power.
The voting quota is set above the average voting quota. You won’t see its effect except if you belong of the top users that vote a lot.
So what are you arguing about ? It will only impact few users, not you.
So you might want to put your comment here over into the original post discussing this new voting feature?
Rude. Thank.
I called it an anti-feature bc voting is a feature, but now this works against that. It’s a
mandatory new feature that removes existing capabilities
You said
The vote quota is the equivalent of scaled feature.
But I believe this is not the case, at least if I am understanding it correctly? Scaled sorting takes the existing votes and sorts them differently than Hot or New, but this new anti-feature will prevent the votes from being allowed to be submitted in the first place. People have already commented that they want to upvote some things - iirc it was a comment of Rimu’s - but could not bc they had already reached their quote for the day.
I doubt I will reach the quota myself personally, but we will ALL be impacted by this change, I believe. It makes PieFed less “welcoming” to discard contributions made to it. It’s a hard enough sell to get people to switch to here from Reddit, without such “features” giving them subtle hints that they aren’t truly wanted here, and reminders that we are resource-starved while in comparison Reddit will bend over backwards to serve their needs (or at least pretend to).
Ask yourself why it was created.
I have asked not only myself but Rimu and everyone else, but I just don’t get it. The various explanations are that it’s “unfair”, and that it will save on resources. However, while it truly is “unfair” that the likes of ThePicardManeuver, PugJesus, etc. have to submit the majority of content across the entire Threadiverse… that is simply how it is right now, and as unbalanced as it is, it seems preferable to limiting their contributions?
For that matter, why don’t we say that it is “unfair” that Rimu makes all of the contributions to the PieFed sourcecode? Should we now act to limit Rimu’s contributions, because others are not keeping up? OF COURSE NOT!!! Nor do we ask our heavy posters to limit the actual posts that they submit…? So how then did votes suddenly become undesirable?
The technical burden argument resonates much more strongly with me. But… Rimu says that it is not for that reason. If it were then I could understand it, but since it is not…
What then is the reason that this is being done? It surely CAN’T be the “unfairness” issue - see above, especially the comparison with Rimu’s contributions - so what then IS the reason for it?
In your example, what if the two people talking the most were the only two adults in the room, and the rest were children? Or what of it were a technical meeting and they were the only two actual developers? Context matters so much there!!! Nothing in life is ever truly going to be “fair”, but at least we all had equal ability to vote on content, before the change, but now we do not - some people have less ability to vote as they wish than others. Consider: someone works very hard all week, then has a single day to check social media, but they reach the cap and then cannot vote anymore - why is their right to do so limited, whereas someone who votes equally but more spread out over the course of a week curtailed? Or someone who works hard all semester long, then has a break after finals week - they want to go on a social media splurge, but cannot, whereas someone who submits 100 times the total amount of votes is allowed to do so, as long as they space it out into multiple days.
You talk as if before was “unfair” - when everyone had the same access to vote as everyone else - but I say that the new method is the truly unfair, when some types of access are favored over others. This is not friendly or welcoming to people who have circumstances that are different than others.
I would be okay with a “rate limiter”, more than a “daily quota”, as the perception of the former is much more fair than the latter.
This will impact me, you, and all of us. If you think more about what I’ve said you will see how. It is one thing to handle existing votes on a different manner (opt-in to doing so), while it is another to mandatorily reject submitted votes right at the source (NOT opt-in, nor opt-out either, except by migration to another instance I guess, if you can figure out how to find one that does differently, and hope that they do not change their minds about the situation later).
And your vote quota is low, i think you won’t be able to reach its limit except if you use a bot to cast vote.
Here is my vote quota after being back since two days.
Yeah few people post lot thing as we do. That’s the same for wikipedia. 90-10-1 rule. We could set a post quota too, i would be fine with reducing the amount of content shared here and slower interaction.
Those example above doesn’t impact the order of your timeline. And yes they are another problem.
I compared it to scaled feature because scaled is balancing our timeline. The voting quota is also balancing our interaction. So for me they are the same tool.
This will impact me, you, and all of us. If you think more about what I’ve said you will see how.
It won’t impact me, i mainly upvote, use chronological order and subcribed. I don’t belong to the top voters. If i reach the quota, i would be fine with it.
Unfortunately, Tournesol association is working on a Bluesky algorithm. But their actual main goal is to make a mathematically good model, they will implement it as a demo on bluesky, but we will be able to use it to work with Mastodon. (The same way the actual Tournesol app for youtube videos is a “demo” of their mathematical model to sort stuff reducing people biases compared to usual notation methods)
As far as I understood, for now, their algorithm uses a mix of Timeline and “Upvotes/Downvotes”. Only that up and downvotes would work differently (Massive downvoters will have less credibility for example). To my eyes, it looks like they are trying to make the “Scaled” filter of Lemmy, with more rigorous mathematicalities inside it
If you go that way, the instance chooser should indicate the vote quota set by admin.
But, honnestly, the vote quota is so high that very few people can reach it. For me, it work as the scaled filter for post.
THIS CHANGE DOES AFFECT YOU.
“It does not affect me directly so I do not care” is not a good argument. e.g. if you drive rather than ride a bicycle to work, then you are still impacted by shutting down or opening up bike lanes - if nothing else by the increased/decreased traffic on the road that you do use.
This change affects us ALL - some users directly in the sense that they cannot offer as many contributions as they used to, while others indirectly by seeing an overall cooling effect with fewer votes overall across all posts, but particularly in niche ones where the impact of just 1-2 extra votes would have made all the difference. Posts that would have hit your Subscribed feed (not sure if you use that or not but it would be affected either way:-) are now going to languish, having no additional votes (beyond the OP’s self one) to boost them there.
And if this change is extended to include comments as well - and why wouldn’t it? How are they so different, really? - then likewise the Active filters will be impacted as well.
This change affects Snoopy in a good way - his votes are no longer being drowned in a flood of indiscriminate votes that water down the effect of votes from more selective people like him.
This doesn’t track. On one side you say this will directly affect incredibly small number of people but this incredibly small number of people also flood posts with their one vote? Wouldn’t that be incredibly small number of votes per post?
Okay but historically, limiting votes does not end well.
So maybe implement ranked choice voting instead? j/k as far as this proposed solution but I hope you see what I mean here in terms of the underlying issue: democracy itself offers both problems and solutions to those who use it, hence why it is failing all across the globe.
Also, at least up until now, people have been offered a full CHOICE as to how much votes affect them, personally. Sorting by New ignores votes entirely, while sorting by Top used to consider all votes, but now will consider only some of them (if half are being discarded, some by the vote quota itself and others by people changing their behavior as a result of needing to constrain their votes, or at least hearing the signal loud and clear that lots and lots of votes are undesirable, hence they stop offering them). I haven’t thought about how this impacts users of Scaled sorting, which I myself never use, personally speaking.
