I just set up my instance and noticed that posts on my Instance dont have the same upvote and comment counts? What couses this and how can I solve it?
Example:
https://lemmy.ohaa.xyz/post/151 - 26 Upvotes and 12 Comments
https://lemmy.ml/post/3331957 - 611 Upvotes and 54 Comments
The only instance that has the true number is the host instance.
other, bigger instances have much more accurate numbers. Thats why Im confused
Basically your instance will start logging the numbers from the first person on your instance interacts with community AND will only log local subscriptions and votes. Obviously a bigger instance is likely to have someone connect before you more than a smaller one. It’s funny because I will create a post from my instance and think it flopped, only to check it from another instance and it did numbers.
The only accurate number is the host instance. You’re not thinking in fediverse terms. What about votes from federated instances? If you have some blocked that they don’t the numbers won’t match. What about banned users? If someone is banned from theirs but not yours the numbers won’t match.
I wonder if we are better off not showing the number of upvotes/downvotes? Like, people can still upvote or downvote, but no one can see them except maybe the instance admins. Maybe collapse highly downvoted comments when they hit a threshold? I know one of the things the fediverse brings is karma doesn’t matter, so it shouldn’t matter especially since it won’t be the same depending on the instance you’re viewing it from. What do you think?
It’s only obvious when users switch between instances, like log out of their account and log into another. In the 5ish months I’ve been here this is the first time I’ve seen it brought up, so to me it’s a non-issue. There are a thousand other more pressing matters
I just set up
just… it does not backfill previous content and votes. What time window are you talking about here?
The whole design is not what I would describe as robust and chock full of features. A lot of people have pushed federation as the key to scalability (now with over 1500 lemmy instances online ) when it is one of the least-mature parts of the code and carries a lot of overhead. Having 500 servers all subscribing to the same communities on lemmy.world for a dozen readers is causing problems with outbound distribution.
What time window are you talking about here?
Ugh… 4 hours. Kinda makes sense. Thanks
