If the community’s server does not track downvotes, then indeed, everyone just sees downvotes by users local to their home instance. Otherwise, downvotes work the same as upvotes - that is, somewhat reliably - unless your frontend or user instance hides them but that’s your problem.
Hell, sometimes even posts don’t federate. There was a period of a few days where Lemmy.World’s federation was messed up, and all my posts here from that period are some of the least-upvoted in this community. I figure those posts just never made it outside of Lemmy.World, or at least past anyone not directly browsing the community.
Like, I thought this post was pretty good, but it only got a handful of votes, and only a few of the comments were from accounts outside Lemmy.World.
You'll see the upvotes from all federated instances. Downvotes are not federated evenly. Kbin is like that, I assume fedia.io is the same. I don't see the downvotes I get on Lemmy instances. Sometimes that's a good thing.
So you only see votes from instances that are federated with that instance.
If instance A is not federated with instance B, but they’re both federated with instance C, looking at a post on instance C from instance A will not show votes from instance B. But looking at it on instance C will show both A and B’s votes.
As for complaints, it’s paranoia. I’ve seen people on .world complain about Hexbear downvoting en mass, and HB doesn’t even have downvotes, and even if they did, .world wouldn’t see them since they’re not federated.
Hexbear brigades with alts on instances which do have downvotes. I am currently running a test instance which attempts to track this (I will post results in a couple of months) and it’s very obvious in some cases that this is happening. There are like 16 or 17 users I’ve found so far which do nothing but downvotes posts engaged with mostly .ml accounts on certain topics.
Ah, I see. As far as I know fedia only defederated a few very small instances that caused technical issues. Hexbear comments were actually one of the examples I was referring to, but upvotes instead of downvotes. But I guess the discrepancy then comes not from my end, but theirs. They probably did not see a lot of the downvotes because they came from instances that weren't federating with them, so they weren't aware of how much they got actually downvoted. Seems a little problematic in that it reinforces their little echochamber.
Also, the synchronization of content between instances is not instantaneous. Even with votes from federated instances, it will take some time for everything to process.
Browsing fedia.io, where my account is. On desktop, Chrome, one of my comments shows one favorite (upvote). On mobile, also Chrome, the same comment shows zero favorites.
I doubt all the vote inaccuracy is only at fedia.io, or other kbin/mbin/forks. Maybe it's unintended, but I also think it's a good thing. Reddit [spit] had to fuzz the votes to prevent spammers and brigading, to what degree that's possible. The inaccurate approval/disapproval also serves to make different voices heard where they would otherwise be more hidden. Getting conversations out of silos and hugboxes is good for society. We all could use some practice interacting with people we don't see eye to eye with.
The post and comment votes do federate, but it’s not particularly reliable. No clue why though. That’s a bit outside my wheelhouse and there is nothing in the documentation about post voting.
Another user explained why I likely saw the discrepancy. Though, aside from that, federation generally is very much wonky and unreliable unfortunately.
Add comment