Rather than communities being hosted by an instance, they should function like hashtags, where each instance hosts posts to that community that originate from their instance, and users viewing the community see the aggregate of all of these. Let me explain.

Currently, communities are created and hosted on a single instance, and are moderated by moderators on that instance. This is generally fine, but it has some undesirable effects:

  • Multiple communities exist for the same topics on different instances, which results in fractured discussions and duplicated posts (as people cross-post the same content to each of them).
  • One moderation team is responsible for all content on that community, meaning that if the moderation team is biased, they can effectively stifle discussion about certain topics.
  • If an instance goes down, even temporarily, all of its communities go down with it.
  • Larger instances tend to edge out similar communities on other instances, which just results in slow consolidation into e.g. lemmy.ml and lemmy.world. This, in turn, puts more strain on their servers and can have performance impact.

I’m proposing a new way of handling this:

  • Rather than visiting a specific community, e.g. worldnews@lemmy.world, you could simply visit the community name, like a hashtag. This is, functionally, the same as visiting that community on your own local instance: [yourinstance]/c/worldnews
    • You’d see posts from all instances (that your instance is aware of), from their individual /worldnews communities, in a single feed.
    • If you create a new post, it would originate from your instance (which effectively would create that community on your instance, if it didn’t previously exist).
    • Other users on other instances would, similarly, see your post in their feed for that “meta community”.
  • Moderation is handled by each instance’s version of that community separately.
    • An instance’s moderators have full moderation rights over all posts, but those moderator actions only apply to that instance’s view of the community.
      • If a post that was posted on lemmy.ml is deleted by a moderator on e.g. lemmy.world, a user viewing the community from lemmy.ml could still see it (unless their moderators had also deleted the post).
      • If a post is deleted by moderators on the instance it was created on, it is effectively deleted for everyone, regardless of instance.
      • This applies to all moderator actions. Banning a user from a community stops them from posting to that instance’s version of the community, and stops their posts from showing up to users viewing the community through that instance.
      • Instances with different worldviews and posting guidelines can co-exist; moderators can curate the view that appears to users on their instance. A user who disagreed with moderator actions could view the community via a different instance instead.
  • Users could still visit the community through another instance, as we do now - in this case, [yourinstance]/c/worldnews@lemmy.world, for example.
    • In this case, you’d see lemmy.world’s “view” of the community, including all of their moderator actions.

The benefit is that communities become decentralized, which is more in line with (my understanding of) the purpose of the fediverse. It stops an instance from becoming large enough to direct discussion on a topic, stops community fragmentation due to multiple versions of the community existing across multiple instances, and makes it easier for smaller communities to pop up (since discoverability is easier - you don’t have to know where a community is hosted, you just need to know the community name, or be able to reasonably guess it. You don’t need to know that a community for e.g. linux exists or where it is, you just need to visit [yourinstance]/c/linux and you’ll see posts.

If an instance wanted to have their own personal version of a community, they could either use a different tag (e.g. world_news instead of worldnews), or, one could choose to view only local posts.

Go ahead, tear me apart and tell me why this is a terrible idea.

  • Max-P@lemmy.max-p.me
    link
    fedilink
    English
    arrow-up
    17
    ·
    3 months ago

    That sounds great and all on paper but that also requires a ton of moderation overhead as now every small instance has to have enough mods to deal with everything being posted, since moderation would be local only. So all the spam and CSAM would have to be taken down by each individual instance. Would also somehow have to find a way for instances to pull the hashtags out of every federated instance too. The way it works on Mastodon is someone follows an account and that causes the data to get pulled in. On Lemmy you don’t follow users, you need a way to pull the data in.

    The end result would be a mess of instances not even agreeing on vote counts with vastly different comments too, and even the posts.

    Lemmy doesn’t aim to be an uncensorable platform. I join communities for the content, the users, and for better or for worse, the mods too.

    The individual problems of having to deal with the duplicate communities will get worked on eventually.

    • KoboldCoterie@pawb.socialOP
      link
      fedilink
      English
      arrow-up
      4
      ·
      3 months ago

      So all the spam and CSAM would have to be taken down by each individual instance.

      Or only by the instance from which they were posted. If an instance is a moderation graveyard and is generating CSAM spam, it probably just needs to be defederated from, but I agree that the necessity to rely on local moderators to cleanly remove a post is a problem with the proposal.

      Would also somehow have to find a way for instances to pull the hashtags out of every federated instance too.

      If each instance shared a list of communities that it hosts with each instance that is aware of it on first discovery and periodically thereafter, it would assist with this. Wouldn’t need to duplicate the content, just share a list of communities that exists. (I think that lack of duplicated content would actually be an improvement over the current system where, unless I’m mistaken, content is being duplicated, but I might also have an imperfect understanding of how it functions now.)