Do we have coolguides yet? This is a good place for it.
this post was submitted on 21 Jun 2023
182 points (99.5% liked)
Lemmy
2172 readers
7 users here now
Everything about Lemmy; bugs, gripes, praises, and advocacy.
For discussion about the lemmy.ml instance, go to [email protected].
founded 4 years ago
MODERATORS
"has anyone from my server interacted or searched for the post by it's URL" is misleading. I struggled with this yesterday. Turns out you have to search in a very specific way.
In both kbin and Lemmy, you can't just go to the community's URL (which is utterly bizarre). You must search the full magazine name. In Lemmy, you weirdly need the ! in front when searching it to find it. In kbin, you don't need that, but you do need to search the magazine in the "neutral" search mode, not magazine search mode (lol wut?). Actually, in Lemmy you also have to use the "normal" search field and not the community search field.
And of course, both have a discovery issue. People want to be able to search a partial string like "hobby" without having to know what instance their community might be on or if the full name might be things like "hobby_discuss", etc. They should not need a separate tool to do this search. That's just a barrier to entry.
Anyway the whole thing is a usability barrier that needs to change. It also makes smaller instances actively harder to use, which is a bad incentive. We don't want people to experience small instances as "buggy" (even if it's working as intended).
Anyone currently trying to create a sub should have an account on every major instance and subscribe to their new sub to ensure it shows up in the search. And yes, that is just completely silly (and unscalable beyond the biggest instances).
In Lemmy, you weirdly need the ! in front when searching it to find it
This hasn't been my experience on Lemmy. I'm regularly able to use for example [email protected] or https://madeup.server/c/anime_tiddies in the search bar and it resolves it both ways. Sometimes you need to wait a few seconds for it to populate though.
I will simply post /c/animetiddies and wait for the server architecture to unbreak itself until my link becomes valid. Since that is obviously the way it would be.
It should have a search function similar to browse feddit which would then add a community, this way it would be much faster and simpler to subscribe to new communities
Thank you. As a newcomer I find this very helpful.
It's dumb that someone needs to interact with community in order for it to federate with your instance. Like, how are you supposed to find it in the first place? It makes it difficult for communities to grow on instances that are not mainstream which makes decentralization useless.
It would be cool if you could see stuff from other servers/instances without someone from your server/instance interacting with it first, yes.
At the moment I am personally not complaining though. I am sure the developers have tons of things they're trying to work on at the moment. And I can still see stuff from other instances/servers.
This is one of the key issues with the entire Fediverse setup and it's not one there are any good answers to.
Not least since it makes being on the biggest instance objectively the best experience, this defeating the point of decentralisation.
In order to avoid this restriction you would need a global instance discovery mechanism, which is extremely hard to implement without a central server that keeps a list of all instances in the network. And if you do implement instance discovery through a central server you really are losing the whole point of decentralization.
Additionally, it's good that each instance does not federate with everyone else by default. If it did, it would have to process all activity and keep a local copy of all the content in the entire network. This would be insanely inefficient, and make it prohibitively expensive to run even a tiny instance with 1 user and no communities.
Decentralization isn't useless if you can't immediately see everything in the network, come on... We're just spoiled by centralized services.
you would need a global instance discovery mechanism
I don't think you do. Instances should merely reach out to other instances it's federated with periodically to get a list of communities and some of their metadata. Ideally, they could ask all of those other instances to notify it when a community is added, modified, or deleted, and then store that metadata.
That should be pretty easy to implement, and maybe it already has, idk.
Sure, but this isn't finding new instances, just new communities on known instances. Indeed, this is not difficult to implement. The reason it's not done already is for resource economy. A lot of instances are already struggling to scale, making them process and store a lot more content with little value for most users of the instance isn't feasible for a lot of servers right now.
A list of communities isn't "a lot more content." Just run it once daily and the problem is solved. Instances don't need to store posts, just community names available on that instance.
If the issue is finding new instances, I think it's fine for it to take some time the first time someone tries to find a community on that instance. But after that, it should immediately have a list of communities from that instance.
Ummmm... All you need is some bots on each instance that automatically will interact once with communities known on lemmyverse.net and boom, you have unlocked full federation for every user on those instances. You are not losing any privacy through that, it just skips the steps where user has to manually index a sublemmy before it federates and makes platform more usable.
Sure, but now this system has a dependency on the "centralized" lemmyverse.net service. And also your instance now has to receive and store a copy of almost the entire network's content. Lots of instances are already struggling to sustain the load, this would make the problem even worse.
If a single instance decides that it can sustain the increased load and doesn't mind depending on lemmyverse.net sure, nothing's stopping them. But it shouldn't be the default behavior for all instances.
If load is that big of a problem then pretty much all instances are in a big trouble because all you need is ONE person that decides to run some bots and it's done anyway.
Well yeah, this is a fair point. It's on the instance admins to vet signups and monitor usage. In general the fediverse has a lot of potential for easy DOS attacks, it's one of the main weak points.
I will send you my post on how we got 550 000 bot accounts in 3 days lol
Oh right, forgot to mention it, lemmyverse.net is open source so there's no problem with centralisation because you can run it yourself 👍
Yet bittorrent dht is 20 years old. How can this supposedly decentralized service be unable to self organize. Is Lemmy some kitchen napkin high school fair project ?
Social media and torrents are pretty damn different. There's a reason no federated platform has implemented automatic discovery, even ones with much more resources than Lemmy, like Mastodon.
I don't know why you folks keep pointing at missing features and saying "Lemmy doesn't have this pretty advanced network feature, so it's not really decentralized", or "it cannot organize", or "it's useless"... It's basically two people's passion project that only blew up in the past month because reddit fucked up. You're not paying for it, are you? So I really don't see how this attitude is warranted.
The tech is right there, it's 20 years old. I'm pointing at it in response to people saying "this is too hard, we can't have 700 instances sharing a few kilobytes of text !, You're asking too much"
load more comments
(2 replies)
I just did some math and assumed there are 700 (some instances are blocking other instances) instances and 12 000 communities. 700*12 000 = 8 400 000, users across the platform need to copy url of community and paste it into search this many times to make the platform fully federate with everything. Numbers were taken from here: https://lemmyverse.net/
wait, so if you federate with another instance through one community, you won't get to see the rest of the feeds from that instance?
would subscribing to the domain directly (like kbin allows, maybe lemmy could in the future) reduce the number of actions to 700^2?
Yes and yes. I have no idea how kbin works but nubers are correct.
Sorry! My bad, numbers are incorrrect but I'm not sure how to calculate correct numbers lol. Wouldn't it be 700x699 or 699x699 because instance wouldn't have to federate with itself?
oh yeah, it is n * (n - 1)
bc ~700 bots need to do ~700 minus 1 actions
a continue inside an if statement checking if the bot domain is the same as the subscribing one would suffice
edit: or use a search algorithm to pop the domain from the stack before starting the loop, that would be more efficient
edit2: even better: assuming all bot names are the same, you just iterate over the stack (constant), and pop each domain (new stack minus popped domain) and feed it to the 2nd loop.
edit3: SCRATCH EVERYTHING. The bot names don't even have to be the same, you just iterate over a constant stack of botname@domain entries and pop the iteration from it... feed the new stack to the second loop (nested) and done. Why did I take so long to reach this?
This would increase server load and traffic by a huge amount though, and would the benefit really be that big?
There are already quite good places to find communities for stuff you're interested in, and less niche stuff will be federated "automatically" by these rules unless you're on a small instance.
Good community lists:
https://lemmyverse.net/communities
https://browse.feddit.de/
Reddit communities now on Lemmy:
https://sub.rehab
https://redditmigration.com/
this number is only completely relevant for someone on an instance all by themselves or with no communities at all. And discounts instances that are or will de-federate either partially or fully. It also assumes some need to be a part of all 12,000 communities. I think tools like you linked solve this issue anyway. I personally believe to a certain extent every community being federated to every instance kind of defeats the purpose of federalization.
But that's not my point though? The point is to federate with all sublemmies of FEDERATED instances. The problem is that when you federate other instance it doesn't federate sublemmies of that instance automatically which limits interactions between instances by a HUGE amount
It's a risk reward question then. That would 100% slowdown the initial federation if it needed to pull in every community and even if that was accepted, should every instance constantly poll any instance it knows about for new communities? Also you aren't guaranteed to need all those instances anyway and then that's just a waste of space and processing power. Correct, it limits interactions but only to what's necessary which allows instances to be ran on lower powered hardware, allowing more people to join in. With the possibility of third party tools I don't see much of an upside of building that into lemmy.
All you need is one person that decides to run a bot for their instance and it will be done anyway.
most people on the internet have way too smooth brain to comprehend this
therefore Lemmy is complex and scary
So, if I setup single-user instance for myself, I will never discover any posts from communities on other servers I’m not subscribed to? Am I reading this graph correctly?
Basically yes. And to subscribe to the communities on the other servers you need to know their full name including the instance name.
Federation is direct communication between websites, so yes, you understand correctly. Sites must learn about each other before posts federate.
Oh, look at Mr. Fancy Pants with his Subscribed tab. If there's a personal front page on kbin, I've not found it yet. :P
wait... so if one of us follows an NSFW, we all get to see boobs? noice!
So what about instances that block other instances? I.e Beehaw. Will they still receive updates?
No, if you're on a blocked instance.
If you're on an instance that isn't blocked, then yes.
Couple questions.
- Is there any pattern for backfilling on a new subscription?
- Direct accessing a post from an unsubscribed server doesn't appear to always load all comments from the host server, is this due to when the content was initially loaded (and maybe some cooldown on refreshing it)?
- Out of curiosity it looks like content is assigned unique ids against the local server, I assume these are internally mapped to their origin server and ids? Is there a way to direct link to correlated posts similar to the
/c/name@serverpattern?/post/id@serverwas my initial guess but doesn't appear to work.
There are a few other interesting things going on in terms of federation speed. This thread looks so different on kbin.social then it does on lemmy.world.
Some of that may be due to defederation but it looks like some of it is federation lag.
view more: next ›