Establishing a norm around karma calculations being opt-in across the fediverse

RIght now lemmy doesn’t calculate or display a user’s “karma”. And many think this a good thing (me included).

Interestingly, kbin does calculate karma, even for us lemmy users (you can all probably just search on kbin.social and find your karma now, +/- federation inconsistencies).

Whenever karma comes up, this fact often comes up, along with the identification of up/down voters, such that many lemmy users will probably know that they actually do have karma and can go look it up if they want to. Some lemmy apps/frontends are also reporting karma AFAIU.

So I think the question now presents itself of whether this is an issue we want users to have some control over, within the bounds of what can done over federation/AP of course.

I can imagine a system where karma is an opt-in setting of one’s profile, and a protocol is established that any platform/client that understands up/down votes ought to respect this setting and that non-compliance risks defederation.

Though lemmy/kbin obviously lean more “public internet resource” than microblogging platforms like mastodon, I think it makes sense to value user health and safety here, and this seems like a not unreasonable option to establish a norm around.

Thoughts?

Coelacanth,
@Coelacanth@feddit.nu avatar

It’s also worth adding that Karma calculation on Lemmy - at least currently - is even more meaningless since it doesn’t work properly. Not only does it fail to calculate your score well [1], deleting a comment resets your total comment score to 0. [2]

[1] Lemmy GitHub issue

[2] Voyager GitHub issue

Kolanaki,
@Kolanaki@yiffit.net avatar

I don’t mind it not being displayed. There is no point for account karma. Shit, even on Reddit where it was used as a value marker for selling an account never made sense to me because the average user doesn’t ever look at a profile and may not even notice your username because they’re kinda small and the focus of the entire site is on the content and not the individual and it wasn’t, like, a thing to think people with tons of karma were more trustworthy. If anything, the opposite was true.

I’ve heard that it meant your posts would have a higher chance of hitting the FP but I think that’s bullshit considering a vast majority of FP posts were not from the same group of people (exception: GallowBoob was everywhere and posted frequently) and often were very low on the karma totem pole.

In smaller communities, you’d see that because there are like 50 people there at any given time and the only like 2 or 3 are posting decent content.

heavyboots,
@heavyboots@lemmy.ml avatar

I kind of feel like karma is pretty much a part of the link aggregator style site’s core usefulness. It’s a quick nod to figuring out if an account is generally well thought of or not and my immediate instinct would pretty much be to block people hiding their karma on here, because it points to an unwillingness to participate in the core voting and being voted on idea.

At the very least, I wouldn’t want accounts that opt out of displaying karma to be able to vote on anything either. All in or stand on the sidelines and watch IMHO.

maegul, (edited )
@maegul@lemmy.ml avatar

pretty much a part of the link aggregator style site’s core usefulness

I mean … lemmy is doing just fine without it, and many here seem to be enjoying its absence, if they have even noticed it, so I’m not sure where you get this from. core usefulness seems like quite a stretch TBH.

The only other places I’ve experienced such a system are reddit and hackernews. On hackernews, its often belittled as irrelevant or something best ignored, IME at least.

On reddit, well, I don’t think much positive came out of karma’s effect there. Let me know if you’ve got counter arguments to that.

I wouldn’t want accounts that opt out of displaying karma to be able to vote on anything either

Interesting.

The main issue I have with this position though is that it seems to ignore and even underemphasise the importance of maintaining good culture through reporting and inter-personal engagement and moderation. If there’s a bad faith actor, instead of relying on karma, I think a better way to go would be establish a culture of people engaging with them and explaining why their behaviour is unacceptable, then reporting if necessary, and moderation suspending/banning them if necessary after that. I’d prefer this because it establishes that maintaining “civility” is in many ways a people problem and not nearly as much of a tech problem as many are inclined to think. It also removes the superficiality of a simple numerical score, which can completely miss context and easily be gamed and stymie open conversaion for fear of being permanently punished by “bad karma”.

Another way of putting this, is to ask by what means does someone earn the right to up/downvote? And I’d say by being a member here that hasn’t been moderated/banned. That’s because the core utility of a vote-score is to more easily sort and assess the feed and post comments and know how the community receives your posts. It’s about the posts first, not the person.

If anyone wants to judge someone by their prior posts/comments, and the votes they got, you are by all means able to browse their history, where their votes are coupled with the context of their posts. But I’d say until they’re moderated/banned, they’re entitled to vote. In the case of a pathologically down voted user (IE, regularly, repeatedly, and substantially downvoted), I think that’s a moderation issue and should be easily captured by moderation tools (it’s like a single SQL query), where again, admins can and should incorporate the context into any decision they make.

shagie,

This is a Lemmy problem - not a fediverse problem.

Other platforms on the fediverse use a like or dislike to mean different things. If you retweet/boost on Mastodon, it’s a like. If you say that you’re not interested in showing up on Mobilizon, that’s a dislike. If you like or dislike something on Peertube, that’s, well, that’s a like or dislike on peer tube.

Suggestion that a Lemmy instance should defederate with these other platforms on the fediverse because they deal with likes and dislikes differently risks isolating Lemmy from the rest of the fediverse.

Suggesting that there is a flag on an individual’s profile that says whether karma is accumulated is one thing. Note that this information isn’t federated out and once it leaves the local instance there’s no way to determine that.

Suggesting that a vote on another Lemmy instance (or platform) on a comment of someone who has disavowed karma on a different instance leads to defederation is a bit of an overreaction.

If you don’t want karma, don’t federate likes or dislikes outside of the instance and don’t accept likes or dislikes coming in. Accept that this will mean that the sort order of hot, top, and active will become less meaningful.

However, the rest of the fediverse doesn’t have this problem.

On “votes aren’t private”. The sooner that people accept that this is the case and this isn’t a centeralized clone of Reddit where one trusts the central authority to manage and protect the secretness of votes, the better. This is Lemmy - not Reddit. It also isn’t the only platform on the Fediverse (and isn’t even 1/10th the size of Mastodon in total user count or monthly average users). Remember that there are other platforms out there and this reads a lot like a new kid on the block trying to set the rules for everyone else in the neighborhood who have been there for a much longer time.

maegul,
@maegul@lemmy.ml avatar

So I’ve responded to some of your points below.

Overall, I have large difficulties with your reply, largely because, I’m sorry to say, I don’t think you understand my proposal and come from a position I’d describe as “tech absolutism” that emphasises technology as the only viable solution and enforces the notion that we all need to accept the limitations of the technology to the point of discounting any value in any other kind of solution. Managing a social media ecosystem is not writing a program. Given your membership on programming.dev, it’s reasonable to conclude that this bias is informed somewhat by what you know and are experienced with, and, equally, what you don’t know and are not experienced with.

In short, I’m not sure you’ve made a single argument against proposal. And, as harsh as this is, until you reply in such a way that makes it clear that you actually understand the proposal and how it would work, I’m unlikely to engage in this discussion.

This is a Lemmy problem - not a fediverse problem.

No. What instances and platforms chose to do with each other’s information is a federation problem. The technical capabilities and limitations of what agreements can or cannot be established by the protocol right now are not the end of the evolution of this technology. Engaging in what federation means at a cultural level is absolutely relevant to evolving the technology and practice. This sort of thing will almost always be a federation problem.

Other platforms on the fediverse use a like or dislike to mean different things. …

Irrelevant and false equivalence. lemmy and kbin, the platforms relevant to this discussion, are almost identical platforms. Also, this isn’t about likes/upvotes, it’s about karma.

Suggesting that there is a flag on an individual’s profile that says whether karma is accumulated is one thing. Note that this information isn’t federated out and once it leaves the local instance there’s no way to determine that.

Yea, this is the proposal, I don’t know why you’re pushing it aside. Also, though karma is an aggregate that, as I understand, is calculated locally on any given instance … you absolutely can tell if the aggregation is done on another instance … just look!

Suggesting that a vote on another Lemmy instance (or platform) on a comment of someone who has disavowed karma on a different instance leads to defederation is a bit of an overreaction …

Not what I’m suggesting at all. This isn’t about votes, it’s about the processing of votes in a particular way and the presentation of the result of that process. Again, fallacies here and at this point I’m not sure you understand what I’m talking about, which is a convention of respecting a user’s expressed wishes on what is displayed about them personally as a condition of federation, not the prohibition of voting in any way. I do not understand how you arrived at this conclusion. At this point, honestly, I’m sorry to say, I’m suspicious of your intentions.

If you don’t want karma, don’t federate likes or dislikes outside of the instance and don’t accept likes or dislikes coming in. Accept that this will mean that the sort order of hot, top, and active will become less meaningful.

This is lazy tech absolutism. It is a position that is neither necessary, obvious, or fruitful.

Remember that there are other platforms out there and this reads a lot like a new kid on the block trying to set the rules for everyone else in the neighborhood who have been there for a much longer time.

Sorry to say … but this made me LOL. Establishing norms over federation isn’t a new idea I’ve just cooked up. Nor is this proposal trying to set any rules … this is about respecting users choices and privacy, which is very much in the spirit of the fediverse. It’s this sort of tech absolutism that is more likely to be the new kid on the block. Moreover, there are no rules, diversity is part of the point. But that doesn’t preclude conventions, expectations and behaviours around establishing and enforcing these. In fact, that’s how you get the diversity, and is part of the point.

density,
density avatar

I am not sure if you saying:

a) assigning karma should be opt-in: all users have 0 karma unless they flip a switch allowing them to accumulate or lose karma, or

b) displaying karma is opt in. in which case do you mean displaying for oneself, for others?

maegul,
@maegul@lemmy.ml avatar

I don’t see a major difference between a and b. The point is that the user can opt in or out of their karma ever being visible to anyone including themselves.

Truth is that the numbers for the calculation are always there, it’s just the sum of vote numbers. But what matters is whether it’s afforded in the UI and therefore whether users and their behaviour are mindful of it.

density,
density avatar

so you mean b

maegul,
@maegul@lemmy.ml avatar

I guess? What was confusing to me is that option a seems quite distinct from what I’m talking about. Karma is always there, as everything has votes. So some option to turn the accumulation of it in and off or to reset it seem like different issues.

density,
density avatar

well, do you think there should be votes at all? why/not?

if you think there oughta be votes, do you think those values should persist outside of the thread they are cast in? I have seen others comment recently to the contrary. the vote exists on the individual statement and goes nowhere.

maegul,
@maegul@lemmy.ml avatar
  • I think votes should exist.
  • They’re a gesture and/or a statement just like replying. They’re terse but often times, as with IRL, all you have to respond with is an affirmation (or disapproval). Mastodon has gone through trying to understand this, where they have the “like”, basically an upvote, but which isn’t used in anyway to affect the feed. Still, people love “liking” posts because of its communicative/social/gestural aspect.
  • Votes are useful for helping someone sort through a feed and tree of comments. Sure it has downsides, such as echo chambers and people aiming for content that gets the most votes, but that’s were separate communities and sorting options come in, though I think a bit more in this regard would work well, some of which is probably coming already. All up, so long as people recognise that maintaining a good culture is important and “up to us”, votes + sorting + communities is a good system IMO.

As for “votes persisting outside of the thread” … I don’t know what you’re referring to, apart, obviously, from user “karma”, though “karma” could be extended to other entities such as communities or instances.

In the case of “user karma”, I’m not absolutely opposed to it, in part because it’s enescapably emergent from upvoting, so any particular platform having it, as in the case of kbin but not lemmy, is always going to happen. I’m in favour of user options and flexibility, as I’ve stated above. Beyond that, I don’t think the idea carries enough merit for it to be the default on a platform. That is, I’d make it “opt in”. I think the only universally understandable purpose for “karma” would be for a person to assess how well they’re being received, and even then, I don’t think that needs to be public.

As I see it, the utility of votes on posts and aggregating user karma are rather distinct, and as the lemmy experience shows, the latter doesn’t naturally follow the former. Many users came to lemmy and almost didn’t notice the lack of karma despite voting and using vote scores to assess and filter posts. Very few have expressed any desire to bring it back. IMO, quantifying a person’s performance, quality etc is just a bad idea.

addie,
@addie@feddit.uk avatar

I think having publicly-viewable ‘karma’ is a mistake - there’s no benefit to having karma-whore posts and bots if it doesn’t gain you anything, and that’s a cancer that eats away at other platforms. Similarly, the effect of brigading a single comment is much less - encourages conversation rather than groupthink. And we can all see each other’s posting history anyway - we can see whether any given account is a troll account and admins can ban them.

Braggston08,
Braggston08 avatar

Kbin user here...

I think karma in relation to up/down vote of posts is a good thing.
The total karma of an account isnt interesting for me, but other users may be more interested in it.
Best solution in my opinion would be if it is handled instance based. If one instance decides to show the total karma of a user its ok for them another instance might decide to not show the total karma or even completely hide up/downvotes.
So every user can find an instance that alligns with his view on this matter.

maegul,
@maegul@lemmy.ml avatar

I appreciate the instance based approach. And it is also probably the easiest from a tech perspective, just as downvotes are optional for an instance.

But it would mean that the user’s wish to not have to worry about their karma would not be respected across the fediverse, and I think that that’s an issue worth considering.

In your proposal, though their karma is not visible on their home instance, it could potentially be visible in every other instance, and therefore of as much concern whether presented on their instance or not.

Crul, (edited )

I’m not an expert, so those who know more, please correct me.

FYI: with the script User Details on Hover you can also see karma in Lemmy. Example.

From what I understand, there are technical issues with this. Allowing people to hide the karma if they want is easy. But blocking people for seeing karma even if they want is much harder. Note that if you want karma for posts and comments (to be able to sort the most voted ones), then the user karma is just a very easy query away (just sum the karma of that user’s posts and comments). EDIT: I realize that this would not solve the issue either: There are technical ways to do anonymous and auditable voting, but I think that would be too overkill for the fediverse.

maegul,
@maegul@lemmy.ml avatar

Thanks!

The contours of this sort of issue on the fediverse often get mistaken, unfortunately. It is very much a “don’t let perfect be the enemy of good” space, where establishing standards and acting to enforce those standards is the MO and relying on tech to provide guarantees without any need of human intervention being very much not the proposed solution.

The actions that people, IE us, can take are establishing norms through dialogue, allowing users to express their own desires, and allowing the federation system to do what it can to find the equilibrium between users’ desires and the amount of connectivity people want.

It can be messy and boring, but that’s what good community management comes down to in the end … people sorting stuff out.

In this particular case, the proposal isn’t to completely prevent a federated instance or a client from doing the calculation and presenting the information, but to get buy-in on the idea that we can “have nice things” without disrespecting people’s needs, and then use our right of association (ie federation) to enforce what we care about.

Sibbo,

You could create a license that makes it illegal to process information about the users in that way. Obviously, someone is gonna violate it, but all the respectable instances will at least not, since they usually don’t operate out of reach of working jurisdictions.

Crul,

IMHO as a random user is that, given the nature of the fediverse, that makes more sense to be an option for instance admins. I’m personally more inclined to leave that decision to each user, but I see how the network effects play a role and how someone would want to enforce their decission on their own instances.

Anyway, it’s an interesting discussion and I like to try to understand the consquences of each implementation.

BuckRowdy,
@BuckRowdy@lemmy.world avatar

Thanks for posting the userscript. I’m seeing lots of RES type features like this and hoping they’ll be rolled into one at some point.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • lemmy@lemmy.ml
  • DreamBathrooms
  • everett
  • tacticalgear
  • magazineikmin
  • thenastyranch
  • rosin
  • tester
  • Youngstown
  • khanakhh
  • slotface
  • ngwrru68w68
  • kavyap
  • mdbf
  • InstantRegret
  • JUstTest
  • osvaldo12
  • GTA5RPClips
  • ethstaker
  • normalnudes
  • Durango
  • cisconetworking
  • anitta
  • modclub
  • cubers
  • Leos
  • provamag3
  • megavids
  • lostlight
  • All magazines
    •