There were also 2 more below that.
And this must be a bot, endless posts by this user, every time the same content on multiple communities.
There were also 2 more below that.
And this must be a bot, endless posts by this user, every time the same content on multiple communities.
Lemmy needs cross communities that exist across instances otherwise it’s going to get more and more fragmented
Why do people insist that there needs to be (for example) /c/politics on every instance? Really, there are only 3 or 4 with any substantial traffic, and there are good reasons to pick one over the others, and they are the same good reasons for them to be separate.
This is a fundamental issue with the way that lemmy is organized that was identified early. Its a design consideration thats pretty much baked into the cake that each lemmy instance effectively tries to be an entire reddit.
Its a bit of an issue, because this will necessarily dilute the kind of network effect that is what allows social media to be as engaging as it can be. Interesting articles don’t get as much momentum. The interaction is more diffuse. Conversations are more spread out and fragmented.
Its beyond the scope of the current design, and I really do commend the developers for what they’ve built (lemmy as is a great experience); however, a more ‘instance’ based approach may have made more sense based on how we’ve seen things scale. Instead of every lemmy instance trying to be ‘all of reddit’ each lemmy instance would focus on a set of niches (for example, a fashion focused instance would have c/fashion, c/mens_fashion_advice, c/streetware,… whatever); then they would federate to propage these niche across the fediverse.
The Star Trek lemmy instance is an example of this. Its a home for all things star trek. I also tried to start something like that focused around WallStreetBets, but afaik, WSB had almost no exodus.
More importantly, the critical mass to get enough users for the content to be interesting didn’t happen. There were too many competing /c’s across the lemmyverse. So articles get posted, but none get more than 1-3 upvotes because the userbase that would get it to say 5-15 upvotes simply isn’t there.
I really do love lemmy for what it is, but this design consideration is absolutely what is preventing Lemmy from being a true Reddit killer. The structure of federation sets lemmy(s) up in a way that there is an inherent blocking factor to super-connectivity.
However, I can imagine a couple solutions to this that dont necessarily require a full burn down and rethinking of lemmy.
One would be to allow for the merging of communities. Users set up C’s, but if there was some way to do a kind of merge (as like on github), where the two RSS feeds could be merged (as in github). Likewise, it would make sense if there were a way to ‘split’ or fork communties, as in, say you’ve got c/fashion, and some one wants to fork off and have it become c/mens_fashion. This would allow communities to consolidate around critical mass (there are enough posts, comments, etc to represent meaningful engagement), and then also to diffuse that issue latter when it makes sense to maybe split off political memes from say, political discussion.
A second solution would be to allow communities to be ‘transferred’ across the federation. This makes sense in that your ‘local’ community should be comprised of the things you care about the most (fashion, mens fashion, streetware, etc…). This feature would allow niche communities to consolidate into single instances, which will also serve to drive engagement (a user of mens fashion is far more likely to post into streetware and vice versus).
A third option would be to build a super structure to lemmy that allows for the consolidation of multiple lemmy RSS feeds into one. Effectively, user would be able to identify various lemmy communities into ‘supper communities’ that consolidate them under a single heading (a tool to grab up all the 'mens fashion advice /c’s across the fediverse) and deliver it in a single RSS feed.
Of the three of these the third option makes the most sense to me. It requires the least rework of the underlying data structures, and seems like a bolt on solution. However, it also might be the least effective of the three. I’ve no intuition about what that would do the structure of the network or if it would aid in overcoming critical thresholds of engagement.
Having instances focused on one specific thing is the best solution, but it requires a couple other problems to be solved first.
The biggest one is discovery. Lets take your example of a fashion instance, hypothetically we’ll call it fashion.world. Lets assume I’m a user interested in fashion setting here on lemmy.world, and I want to subscribe to a fashion community. Currently the lowest resistance method is to hop over to the local community list and scroll through looking for any fashion related communities. If I’m a little more savvy maybe I hop over to the search option and take a crack at some plausible sounding community names starting with just fashion. That might work, but it relies on lemmy.world already being federated with fashion.world, which in turn relies on another user having already found and subscribed to one of their communities. On a very large instance like this one that’s probably a decent chance of having occurred, but on small or obscure instances it’s very unlikely. So we have a massive discoverability problem now. There needs to be some kind of centralized registry where you can type a term and see all the communities across all the instances that might be related to that term.
Another related problem is that instances, communities, and users, are closely bound to each other. I think it was a mistake to put everything together. It simplified things in the early days, it makes it possible to treat a lemmy instance as a mini-reddit, but it causes problems in the long run. Instead you should have a service for users to authenticate with and federate user messages and such, and an instance for communities to be hosted out of and federated. This would also simplify some aspects of moderation as user instances could setup a consistent set of rules they expect their users to follow. If you get caught not following those rules you get banned from the instance. Communities then could have their own rules they setup and via de-federation with different user registries you’d have a quick way of deciding the kinds of users you want in your community. Seeing a lot of hate speech coming out of the user registry run like a 4chan board? Sorry fellas, ban hammer time, that’s why you can’t have nice things. Not to mention breaking users and communities apart lets things scale in a more natural fashion, where the load the community server is under is directly proportional to the interest in those communities rather than if that instance happened to be the most well known one when someone went to register their account.
Before I knew about Lemmy I was in the process of designing my own replacement platform. While I think that decentralization is good, I felt like it should be done at the community level. So everyone is federated to a few user account instances, then each community is a self-hosted instance.
Obviously we’re too far past that point to do it with Lemmy, but it does feel like federation can be an obstacle as much as it can be a benefit
when admins leavy defederation like a nuclear weapon, you know its a problem.
Just bc there are 100 things called /c/whatever doesn’t mean people will fragment. On Reddit, the only difference is the names would have to be different in a new way. You can quickly sort by active, find the active community, and use it. The others will die, and the network effect will live.
I just did a massive cleanup now that I’ve been on Lemmy long enough to know which communities on which instances are active. I’ve had to do this on Reddit as well.