972
If everyone was spread out onto different instances, and communities were based all over the fediverse, the decisions of one instance would be less impactful.
(media.kbin.cafe)
this post was submitted on 15 Aug 2023
972 points (96.2% liked)
Memes
45158 readers
3310 users here now
Rules:
- Be civil and nice.
- Try not to excessively repost, as a rule of thumb, wait at least 2 months to do it if you have to.
founded 5 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
It kinda makes me wish that instances were forced to be single-topic, or even single-community, and that authentication was key-based so that you didn't need to "make" an account on a single instance.
The key based (and content addressing based) thing is what bluesky is building. They're starting of with Twitterish microblogging, but there's people building forums on top the protocol too. Federated, of course.
Lemmy is built upon the ActivityPub Protocol which has the flaws mentioned above Bluesky is built upon the AtProtocol which to me also looks kinda great https://yewtu.be/watch?v=wJBCpzM1VfM ;- Video that explains the difference (i just watched it minutes ago) https://atproto.com/docs ;-Docs for learning how it works
User accounts being key-based/portable is one of the strengths of the nostr protocol and Bluesky/AT Protocol.
I think instead instances should have every community. There isn't one /c/books, every server has a /c/books. Your feed pages just pulls from the entire fediverse. No concept of "creating" /c/books, it just is.
Likewise, there isn't "a" moderator. Every user is a moderator. Whether you vote, or delete the post out ban the user (from your view), your moderation opinions are published publicly. Your local feed algorithm sees everyone's "moderation opinions", if the consensus of the community is delete, then it just doesn't show up in your thread
For each "moderation opinions" by a user, your client investigates their historical record to address credibility and likelyness of being a bot, a user's history is his credibility
I've got similar ideas, but not entirely the same.
What you call communities would be closer to what I would call content sources / repositories (host servers) plus topic tags. Then instead of consensus (because that's too hard to automate with decent quality results) you'd have communities formed by subscribing to "curation feeds" which pull submissions and comment from all over the network in a similar style.
This would let you easily crosspost and comment to multiple related communities in a network, as well as to yeet bad mods/curators without losing any content or splitting the community (just create a new curation feed and get people to switch). You could similarly choose to have your client mix comment from multiple curation feeds (similar to "multireddits" on reddit).
Whatever the solution, it needs to create communal view of content or else users will not have a communal experience of which is the basis for a community. This is why multireddit remained a niche feature incapable of overcoming zealous moderation and censorship.
As a midpoint there's things you can do like "2/3 consensus of X, Y and Z's submission selections on topics ABC", then defining that as it's own feed people can subscribe to.
But it gets complicated to mix and match when different subcommunities have very different local cultures.
Usenet news - 2020s edition
I wish