I agree with the premise of the article. I’m a new Mastodon user trying to give the platform an honest shot, but I can’t understand why I would choose any particular server over another. Why choose anything other than the pseudo-default Mastodon.social? If I’m forced to choose another server, is that choice actually meaningful? I feel like I’m randomly picking between Hachyderm and TechHub.io and others. One advantage of federation I keep hearing is that servers can block other servers if their users become problematic. My question is this: why would all problematic users congregate on the same server? Isn’t it far more likely that you’ll have to block individual users across numerous servers to keep the platform sane?
I do see the potential advantages of federation to prevent the same fate that overcame Twitter and Reddit, but I can’t figure out how to make a user’s server choice actually meaningful. Without solving that piece of the puzzle, I don’t see any practical advantage to federated services.
Problematic users do tend to congregate on certain servers because they keep getting banned from better moderated ones. Eventually they find the ones that won't ban them, and now well moderated instances have an easy target to block them wholesale. That's how you end up with instances like explodingheads.
It doesn't really matter whether the entire world agrees, only that each community comes to an agreement. Mastodon.social collectively agrees on one definition that excludes explodingheads that involves one view on hate speech and spam, and explodingheads collectively agrees on another definition that excludes mastodon.social and involves censorship and free speech. Both groups are happy.
This is the point of federation though. It allows for different standards of moderation to actually exist. If you don't like the way a server moderates, then don't join it (or transfer to another server, which you can mostly do with Mastodon). Hell, run your own mastodon server (or use a "mastodon-as-a-service" service) which is pretty trivial for micro instances.
The idea that there is one globally correct way to moderate is demonstratably completely untrue, regardless of what your actual views are. There must be multiple moderation standards that users can pick from.
The only acceptable approach to moderation is one you can opt out of. Don't want a spam filter - fine, you're on your own.
Having to choose between politically charged "you can't read that" gatekeepers is a dystopian nightmare. Yes you can make your own, but next thing you know is "block this list of servers or we block you".
I don't get why some people are so anxious to block alleged hate speech and supposed misinformation. Social networking sites already allow users to choose their own community simply by chosing whom they follow. Without federation. Why do we need additional segregation and polarization beyond that?
Existing social networks do a poor job though of actually allowing users to choose their own community. For Twitter specifically algorthmic timeline, replies, likes, and retweets will often put content from people you don't follow right infront of you. You can maintain your timeline more by unfollowing and blocking, but this is a bunch of effort you must do on a more individual level.
Mastodon's approach allows you to outsource a part of that by just joining an instance that doesn't even allow "alleged hate speech" into your timeline. That's the appeal of that.
But the point of federated Mastodon is not to have a platform that is solely moderated less than Twitter.
It's that:
- it's not twitter (or rather, it's a "twitter clone" that's non-commercial, not reliant on a single entity, has a degree of data portability built in - decentralised)
- it can allow for the spectrum of moderation styles to exist. The majority of bigger "mainstream" instances are probably going to opt to moderate similarly-ish to social networks, but this doesn't negate the other reasons for mastodon to exist. You can have instances that have an even stricter moderation standard, or even looser.
But at the end you don't need to get why one person prefers a network in one way, because federation means different ways can happily exist, and you chose what you prefer. Just don't be surprised if other people aren't enthusiastic about the people you hang out with.
> For Twitter the algorthmic timeline means who you follow has such a small impact on who and what you see. And even if you switch to the non-default cronological timeline,
"Non-default" sounds too harsh here. It's literally a big tab saying "Following". It's only default for the very first time you log into Twitter, afterward it will remember your decision (it will be "default").
> , replies, likes, and retweets will often put content from people you don't follow into your timeline.
No only retweets will show in the Following tab. Which seems reasonable, as retweets were historically done via copy&paste. The other things aren't the case anymore for the "following" tab since Musk introduced it. (By the way, after pg himself asked him to add that possibility!)
I agree that Facebook is way worse in terms of moderation, but Mastodon doesn't seem an improvement over Twitter, quite the opposite. Those moderators are moderating for users, when this should be simply left to their decision of whom to follow. People or posts will be hidden from users who would not have hidden those themselves.
The decentralized approach sounds cool initially, but it leads only to even more segregation / filter bubbles / echo chambers. As if the choice of whom you follow didn't cause enough of that.
I think that's a real issue. Political polarization increases since the amount of media channels increases. When there were only a handful of newspapers or TV channels, they couldn't be too politically biased, otherwise they would scare away most other people. This changed when the number of media outlets increased. And with social media, every user now is able to create their own little reinforcement chamber.
You're acting like the current methods aren't greeted with the same cries.
To answer your question; because they are demonstrably required. Different subreddits already moderate to different standards - why do you insist that each community needs to fit in as part of the same whole?
Because Reddit isn't a social networking site where you create your own community by following people. In Reddit other people create those communities, because Reddit is a forum.
I might talk about allegations and suppositions about an individual's behavior until I've seen evidence, but the hate and disinfo groups aren't imaginary.
Groups that promote hate speech and misinformation do so by invading other communities and being disruptive. Brigading and doxxing have been their tools for a long time.
Personally, I've seen that breaking up groups that promote hate speech and misinformation helps the remaining members to enjoy the community in peace.
The problem is that what is considered "hate speech" and "misinformation" is often very subjective. How would someone even "invade" other people's communities? Everyone chooses their own timeline after all.
If all you ever do is read and never post anything people might interact with, then that's true. That's not much of a "community" though if nobody ever posts.
"Interact with"? You mean reply? If your tweet goes viral, you might indeed get unwanted replies. But usually you only get them from people who follow you, or other people close in your circle.
I mean if that opinion was they can send commercial messages to unsolicited users I don't have to imagine at all. I've been on the internet long enough to remember the days before spam filters.
Ideas are like that, paper can't blush. For a salient example, the idea of providing 12 year olds with puberty blockers and opposite sex hormones without their parents' consent didn't quite fly then, and today disagreeing with it puts you in quite a few naughty lists.
"Ideas are controversial" is no where near "ideas weren't fathomable". There are several 24 news networks that make your exact example of "unallowed opinion" as a seemingly huge percentage of their content.
But also, that's a complete strawman. No serious group I am advocates medical treatment without parental consent. There is a strong movement to let 12-year-old choose what pronouns to use in school even without parental consent.
It's more that at the time you could disagree on some things where now only one view is acceptable. The Overton window has narrowed substantially on certain subjects.
> ... but I can’t understand why I would choose any particular server over another.
Email is often a good analogy for questions like these: Why do you pick one email service over the other? Maybe you think the domain is cool. Maybe you heard that they are reliable or have features that others don't. Maybe your university, company or other community you belong to runs its own service. Or you decide to host your own. Or you really have no preference and you just go with the one your friend uses.
We just don't have just don't have the gmail-equivalent catch-all choices yet. Maybe that will come later or maybe it won't.
Except that with e-mail there are obvious good reasons, which basically boil down to whether you've centered your digital life around the Google, Microsoft, or Apple ecosystems (or never moved away from Yahoo, etc.).
These are recognizably proven brands that aren't going to disappear next year, and which integrate with your other tools in ways that clearly provide benefit.
And back in the day before webmail was a thing, your ISP gave you an e-mail address so you didn't pick at all.
But none of these are even remotely applicable to Mastodon instances. So I'm not sure that e-mail is actually a helpful analogy at all here.
Maybe not from the outside. But after a few months of actively using the platform with a curious mindset, I'd wager you'd have a much stronger idea of how instances differ. That's the main service I see the "pseudo-default" mastodon.social providing in the current world—as a "jumping off" point that people can use prior to migrating to more specific instances with more customizations or better communities.
That's what I meant by "no catch-all choices yet". There is no reason Google or Apple couldn't get in the game if they choose to. Supposedly Facebook is making a go at it for whatever that's worth.
I offered the email analogy because that helped me to wrap my head around it when I was starting out.
Email is much better than that. Even if you don't want to maintain your own email server you can still register your own domain and assign it to a hosting service for a trivial amount.
You can also do that with Mastodon if you're keen, so email analogy holds.
And for both email and the Fediverse, there's an ideal answer that requires only marginally more work: have your own domain. In both cases, you don't have to host a server yourself, you can pay for email hosting and use your own domain. For email, services like Fastmail let you bring your own domain. For the Fediverse, https://togethr.party/ works well.
This works with email, but not so much with Fediverse. There is no standard or - AFAIK - even attempts domain/account sharing (unless you hack your way around), you'll have to be you@mastodon.your.name and you@lemmy.your.name and you@peertube.your.name, but not [at least not easily] you@your.name if you want to use it with multiple systems.
I had absolutely no issues with hosting my own mastodon, I pay $7 dollars a month for a VPS with backups. I'm the only user that can register on my instance so it's just me.
I'm `@aj@s.aj.immo` and I can follow and have conversations with other mastodon users of various servers seamlessly. I was just doing it this evening in a thread while tagging four different people on four different servers. When people tag me in posts it just looks like `@aj`
This is the right way to do this stuff in my opinion.
> I had absolutely no issues with hosting my own mastodon
Just Mastodon - it's not what I'm talking about. Please let me know if my parent comment wasn't clear enough.
Mastodon certainly works and you can be just @aj (as long as no one else with the same host-local part walks in, I guess) no matter where the actual conversation is hosted. But what about Lemmy, Peertube, or other Fediverse systems? I'm aware that Mastodon can talk to them with some success, but AFAIK it is only optimized to present certain types of content, so if you want Reddit-like user experience rather than Twitter-like one you want Lenny. And then you need a different subdomain. And then you realize you're splitting your identity on the same network in two - and in my opinion that's an indicator that something is wrong.
>I'm aware that Mastodon can talk to them with some success, but AFAIK it is only optimized to present certain types of content, so if you want Reddit-like user experience rather than Twitter-like one you want Lenny.
This is partially true. I run Pleroma, which is a microblogging server like Mastodon, so Mastodon content that federates to my instance shows up identically to my own posts. On a system like Lemmy, when their posts federate to me, what I see natively on my instance is a person posting a link that links to the originating website. I can see some replies inline, but they're linear and not threaded like on Lemmy. However, unlike what you said, still using my @me@my.domain identity, I can click the link from my instance to the Lemmy instance and interact with the content there natively. No need to create a different subdomain. There's no need to host a Lemmy instance yourself if you want to interact with Lemmy instances. Lemmy on the backend is still talking the ActivityPub protocol and it's only the frontend that alters how the posts are presented, so as long as your server speaks ActivityPub, you can use the same identity.
Are you saying that the primary issue is that for each federated service we need a separate account / instance for it as opposed to email which is just one email account? I don't disagree with that but it's the same situation for Reddit, YouTube, Twitter. I have to maintain separate accounts on those platforms too.
Yes. And I get it that we all have separate accounts. The difference is they're entirely separate networks, not one. If I'm on Fediverse, and self-hosting, with my own domain, it's fairly natural to want to have a solid identity, isn't it?
Which is where it's different to email. When self-hosting mail, you can have separate identities too, but those will be created on a purpose (different addresses for different personas/purposes), rather than fragmented because of necessity.
But you don't have separate identities for Youtube, Gmail, Meet et.c. For a federated internet, you should have the same ID for all federated services so that you can build the same product experience as with Apple or Microsoft or Google.
It'd be nice if there were a standard forwarding-interoperability server such that several services could share the same account name and local copies of the messages to/from that account. But in the meantime, having @you@your.domain be a Fediverse service like Pleroma or Mastodon, and then having @you@service.your.domain be the handle for each specialized service, still seems preferable to using a third-party server. And if in the future some mechanism exists for those services to share, you would then be able to use the standard redirect mechanism to preserve the old names while using the shared top-level one as the canonical account name.
Your email choice used to reflect your ISP, and now reflects your chosen platform for file storage, image albums, productivity suite, calendaring, etc.
I'm not sure how this applies to Mastodon servers in practice.
The same way people manage to decide whether to buy buns at the bakery on the right rather than the one on the left or donate to the NGO1 rather than NGO2 doing exactly the same, you ought to manage to judge which instance is a better fit or if it even matters in your case.
>why would all problematic users congregate on the same server? Isn’t it far more likely that you’ll have to block individual users across numerous servers to keep the platform sane?
Well, exactly. And who's doing all the blocking and filtering? What are the rules? And which community is better at reporting and keeping the instance the way you like it? What about your interests? If you join a busy generic instance, of course, you will become disinterested in your local timeline.
But why force people to categorize themselves? There's no special type of Twitter for people who are musicians, for example, or people who do airbrush paintings of squirrels. Who knows what you might want to send messages about tomorrow or next year or in half an hour?
Having "themed" instances implies that you're committing to whatever that theme is, but all you want to do is share whatever is on your mind from day to day. This apparent pigeonholing is enough to make the whole idea a failure. Remember Yahoo? It wanted you to drill down through dozens or hundreds of canned search categories... and then Alta Vista came along with just a text box. And of course today that's Google too.
Speaking of search... the bizarre and rabid animosity expressed by many Mastodon users toward FINDING CONTENT on Mastodon is baffling. I've seen threads where someone asks how to search for things on Mastodon and gets berated for even suggesting that people who PUBLISH STUFF on the Internet would ever want it found.
No one's forcing people to categorize themselves. You can be on a non-specific server, or you can be on one for a particular community—but even then, nothing precludes you from talking about other things. It may be the case that instance pickers/lists/descriptions aren't clear enough that you don't _have_ to limit yourself, but there's nothing forced about it.
The reality is that every server will have some local version of a bunch of the most popular things, and that will eventually reach a steady state with a tail. A few big ones, more medium size ones, decent number of small ones...
There's no point denying it, discoverability will be harder and communities will be smaller. So if you actually know of a really good community or magazine to point someone to, just do that. They can read the content without logging in. If they want to participate, then the server that they found a community they want to participate in is the obvious place to create an account, from which they can see everything else as you've said.
Little is being forced by the choice of server/instance, but that's not helping solve the hard part. The hard part is them finding what they need to find. They need a librarian, not a tech support desk. Tech people constantly assume the solution is technical... make the popular destinations popular by talking about them everywhere.
> The same way people manage to decide whether to buy buns at the bakery on the right rather than the one on the left or donate to the NGO1 rather than NGO2 doing exactly the same, you ought to manage to judge which instance is a better fit or if it even matters in your case.
Bakeries and NGOs aren't sticky. Your relationships with them is a series of one-time transactions, and at any time you can switch to the alternative at no cost.
Picking a Mastodon instance is more like picking a school or university - it's a choice of where to commit, made at a point when you're least equipped to make a good call, and increasingly hard to reverse the longer you go along with it.
Perhaps Mastodon is just too immature at this stage for the servers to really differentiate themselves from each other. However, as the number of users and posts on Mastodon grows, will any server actually be capable of adequate moderation? Twitter has ~6,000 tweets per second—it's hard to imagine any Mastodon server today being able to handle even a fraction of that content.
It would be interesting to know how many moderators Twitter has. They do have ad income, so they can afford to employ full time staff. Then again, Reddit does well with volunteer moderators.
I have also noticed that in some cases....user blocks are being replicated to an instance I run. I have this huge and ever growing list of banned users on my single user instance in lemmy. ANd I have no idea who is banning and why...I doesnt seem to affect me...but its odd
Which is exactly what worries me. I don’t want other people deciding who I can read and who I can’t. I don’t want other people deciding what voices should be heard.
This just seems like the creation of another echo chamber.
Well, turn the automagic updating of the banlists off then. It's configurable, but it's not the default because most people are not interested in seeing the posts of (as an example) people who upload kiddie porn, nor do they want that on their hard drive in the off chance the FBI comes by with questions.
So the possibility of highly illegal content justifies a block list which perhaps includes mainly blocking unwanted political opinions? That argument sounds familiar.
>I have also noticed that in some cases....user blocks are being replicated to an instance I run. I have this huge and ever growing list of banned users on my single user instance in lemmy. ANd I have no idea who is banning and why...I doesnt seem to affect me...but its odd
IIUC (and I may not) there are blocklists that are shared between fediverse instances that, depending on the configuration, are automagically updated/applied on a regular basis.
Perhaps that's what's going on with your instance?
Edit: Modified comment to address GP's specific issue.
> why would all problematic users congregate on the same server? Isn’t it far more likely that you’ll have to block individual users across numerous servers to keep the platform sane?
The fediverse is not one unified whole, the fediverse is a bunch of loosely-connected communities that can very easily throw up walls at a moment's notice.
Admins are actively incentivised to kick problem users off. If a server becomes known as a source of problems, then other admins will defederate from it. How you define "problems" depends on who you are, in general if an admin looks at a report for one of your users and finds you doing the exact same kind of bullshit as what your user was reported for then they will probably defed from you without a second thought. After a while your server will only be federated with other servers that are also fine with that particular kind of bullshit. If your particular kind of bullshit is "finding people who are different from you and making fun of them" then there's a high chance that the users on your server will either wander away, or turn on each other; both of these tend to lead to admins deciding it's not worth keeping the place running in my experience.
Mastodon.social is way too big to realistically be handled by anything but a bigger crew of paid moderators than it has, and "we should defed from mastodon.social because it's right on the border of being a constant source of trouble" is a topic that admins constantly chew on in places like the #fediblock tag. But you can start on one of the big servers and move, too. Possibly to a small one with a lot of your friends, so that the "local" timeline is useful. Possibly one run by one of your friends, who will make largely the same "oh god look at this server full of assholes, block'd" calls that you would, and save you the trouble. Possibly one with a cool name, would you rather be sxg@mastodon.social, sxg@possums.gay, sxg@gop.party, sxg@...?
(As far as I know neither Lemmy or Kbin offer account migration yet, so you're stuck on your initial one until they fix this. I'm sure it's in their roadmaps.)
At least as big an issue for Lemmy and Kbin is community migration. A server that hosts a popular community shutting down is a very real risk, and simply telling users to switch doesn't scale. (No, not everything needs to scale, but it would be better if that did.)
I just did a quick search, and users have opened issues asking for both user and community migration on both. So they're certainly aware of it. When it will be added, I dunno - it took a while for Mastodon to get user migration, and I wouldn't be surprised if it turns out that community migration is a larger problem than user migration.
Lemmy seems to have a lot more contributors than Kbin, but who knows, maybe someone will dust off their god-tier 100x PHP skills and Kbin will get it first.
What you see on Mastodon is (a) stuff posted by other people on your server, and (b) stuff posted by people on other servers whom people on your server are following. So your choice of server has a strong affect on your feed. And hosting your own instance means that you really need to start following loads of people if you want to keep an eye on what's going on. I suppose I can see this as a good thing, or at least vaguely see why other people might see it as a good thing, but I mostly find it annoying. (It also means that there isn't an overall Mastodon zeitgeist, the way there may have been with Twitter.)
Another, even more annoying side-effect, is that conversations may have bits silently missing, if they're posted by someone not on your instance. I have been reading a thread of seven posts, posted by five different people, and then looked at that same thread from another server and found that it was actually sixteen posts, posted by eight different people, several of which had not federated to my server. I have been irritated by two different people asking me the same questions, leading to two separate threads hanging off my initial comment, covering exactly the same subject, until I looked at the conversations from their servers and realised that they were invisible to each other.
I would never use Mastodon for anything serious: the chance of irritating people, missing context, and talking at cross purposes are far too high.
>I can’t understand why I would choose any particular server over another
Who runs the server? On what kind of funding, and how long can it last? Are you served intrusive ads? Are your a customer or a product for the people running this instance? How hard is it to run an instance yourself, and what happens if you suddenly post a video that attracts 10M views?
So many questions, I know, that's too hard, gimme something easy.
At the time I signed up for Mastodon and a server, I didn't consider and wasn't presented with any of these questions/answers or other points of differentiation among servers. I guess that's part of the issue: when an ordinary user signs up for Mastodon, they're unaware of what they should even consider when picking a server. As an ordinary Mastodon user myself, I just wanted to get on the platform and start using it.
Well signing up for anything important (or anything at all really) you should consider all of that, federated or not (especially if not). Yes, it's a problem that the majority of people "just want to start using it" and wouldn't consider any of the above when exposing, I assume, their lives to strangers.
Sure, on paper we should all carefully consider every decision we make. However, in practice I’m confident the vast majority of Mastodon and Twitter users (including myself) don’t consider using either service particularly important or deserving of more than a couple minutes of thought. From Mastodon, I aim to follow some interesting users to find interesting articles, learn new things, get a sense of “what’s happening” in the circles I follow, and broadly be entertained for a few minutes at a time. I absolutely don’t consider it a critical component of my life and can’t justify spending much thought on it.
>but I can’t figure out how to make a user’s server choice actually meaningful
in theory, the instance-specific bit of it is the moderation. the one you sign up with chooses which other instances to federate with, as well as how to moderate the content on their own instance. in theory, you could sign up for an instance that only allowed the sort of politics you agree with, or that had very lax moderation, or that had very strict moderation.
unfortunately, none of the fediverse seems to actually talk about this so without signing up and using an instance for a while to see what the moderation policies are like there's no way to know.
> unfortunately, none of the fediverse seems to actually talk about this so without signing up and using an instance for a while to see what the moderation policies are like there's no way to know.
Correct, but that's why you can actually migrate your account (including your followers!) from one instance another. The initial choice doesn't matter too much. Just migrate if you don't like it. That's why there isn't much discussions about it.
"Defederated" is not something that happens to an instance, it's a choice that other instances make. There's no "globally defederated" state. If your server (Server A) is blocked by Server B, then no, you can't move to Server B. But you can still move to Server C that doesn't block you.
Well so this is kind of what I was getting at. It feels similar to an email service being blocklisted, only in this case, you really _can’t_ move your content because at the end of the day, a group of federated instances will also be wary of instances without reputation, as much as those of ill repute. So no, in reality, you can’t move your content, or if you can right now it’s a foible of the system as it is today and will not continue as such.
The phone network is a federated service. You can call an ATT customer from Verizon. You can also call them from Joe's Discount Boxes and VoIP sales.
Now, if you start making spammy calls, ATT might block you as a customer (after warning you first). Same for Verizon. But Joe's Discount Boxes and VoIP sales doesn't, because box sales are down and these VoIP margins are so, so juicy.
Spammy customers flock to Joe, because he allows it. Joe keeps letting customers on, without doing due diligence, without caring if they're legit (beyond a valid looking credit card), and eventually gets blocked by FCC order.
(See https://docs.fcc.gov/public/attachments/DA-23-389A1.pdf, for an example).
Now, federation doesn't have an FCC, the various server admins will be filling that role themselves, but there is a reason scummy people will use scummy servers.
Side note: This kind of FCC enforcement is fairly new to the industry, so if you're wondering why you get spam called all the time still, there is some hope this'll change.
Yes, all problematic users don’t congregate on the same server. Problematic users can pop up anywhere. However, a server could fail to moderate effectively and become a source of misinformation and trolling. That’s the benefit of federation: it ensures a minimum level of moderation quality. Servers that don’t moderate are blocked, which sets up the proper incentives.
Sure, but as the number of users on the federated service grows, how do we not end up in the same place as email today? Does it not become good moderation policy to auto-block all new and small servers (e.g., self-hosted email servers) if you're a big server (e.g., Gmail)? If that happens, wouldn't all users migrate to the big server (e.g., Gmail) and effectively re-establish centralization and undermine federation?
Email isn't fully centralized, and the centralization seems to have plateaued. Fully hosting your own email server is difficult, but there are hundreds of non-Google/Microsoft/Apple hosts. If the Fediverse centralizes to the extent of email, I still prefer that to the current 100% siloization of all social media.
Email is also old and unextendable, whereas the Fediverse has interoperable stacks built on top of it, opening the door to better moderation and web-of-trust tools that help combat spam and decrease the pressure for centralization.
The inverse happened recently on Lemmy. Beehaw.org, a community focused on good behavior, defederated lemmy.world and another smaller site because those sites caused a lot of unwanted content and moderator intervention. Lemmygrad.ml, a bunch of communist extremists, also gets blocked quite often, not just by principled servers like beehaw.
A lot of people got mad for being blocked, but honestly I can't blame the admins for trying to reduce the amount of moderation necessary. There are thousands of people per moderator on those sites and that combined with open registration is a recipe for toxic bullshit nobody wants to deal with.
Smaller instances don't get blocked, though. I doubt they will be until the low quality shitposters will figure out a way to host a server of their own, but without a central gathering place I think even then the abuse will be easily prevented.
That said, Lemmy used to be set to whitelist-only by default for ages. This has been changed, but if someone wants to federate with just their servers of choice, they can just toggle a setting.
Lemmy also comes with an approval form by default, which is used manually to vet users during signup. Open signup servers disable that field, obviously, but those servers are often also the ones that get blocked.
I can see this evolving to a situation where only closed or manually approved servers get federated by default. Requiring manual reduces the probability that a small team of admins will get overwhelmed by a large influx of users.
What's misinformation? Things you don't think are true? Things the instance admin doesn't think are true? Things officially deemed untrue by the State? Or by the UN/WHO? Is is considered Misinformation if it was deemed Untrue in the past by State Officials but is now Official State Policy? We have always been at war with Eastasia. We have always recommended the wearing of masks. We have always seriously considered the possibility of a lab leak. Border closures are a bad, racist, problematic idea. No wait, now New Zealand is the hero of the pandemic because it closed its borders. We have always recommended border closures!
Servers defederating because of viagra/nigerian prince/etc. spam is one thing. A "lefty politics" server defederating with a "righty politics" server or vice-versa because they just want an echo chamber? Sure, you do you. General-purpose defederating other general-purpose servers because of "misinformation" and "trolling", when those terms basically mean "disagreeing with the current mainstream narrative on any controversial issue"? That's absurd. The generally-accepted issue with social media was for years that it created echo chambers. Now people are trying to use mastodon to create even more powerful echo chambers.
There seems to be an assumption in your writing that people should use social media to challenge themselves, be exposed to different views, to "discourse" and so on. Nearly everyone I know via social media wants to hang out with their friends, not debate people. Whether something is an echo chamber is sort of beside their goal, they want to have fun.
If one seeks out something that explicitly isn't an echo chamber, I'm sure there are many places for that. I'm just not seeing that as, really, at all desirable for a lot of people.
The problem is that Mastodon puts a second filter bubble on top of the old who-do-I-follow filter bubble. Except this additional filter bubble isn't controlled by yourself.
There is always a second filter bubble. In the case of non-federated sites, that filter is in the form of the policy of the site and also the algorithms used to sort feeds and replies.
They have a choice? Is it transparent for average users what is withheld from them and where? And where their own posts are withheld? If no, they don't really have a choice. If such a choice even exists.
Twitter is fairly lightly moderated I think, and the "following" tab gives a raw timeline.
Mass federation is the best solution to resist mass surveillance and censorship. I never have a problem picking an instance because I just create my own instance whenever I want to use different SW. In a few minutes I can usually spin one up. and I have a few running on my SBC sitting here on my desk. I rarely ever block anyone as unfollowing is good enough.
Current federation looks like that, a bunch of small independent communities focused on a single topic like tech or photography or whatever, but it doesn't have to be like that. If federation catches on then the most likely the pick will be between 5-10 huge general instances run by corporations.
In such world, Twitter and Facebook could add support for ActivityPub, Google and Microsoft could build their own, different ISPs and media companies could run their instances, and so on. Then the benefit of choice becomes more clear - who do you trust the most to provide you an account. Basically, it would be like the situation with email where most people use a 3-4 biggest services, but the ecosystem is open to new competitors.
The current state of federated social networks is very experimental. If it catches on with big corporations 90% of people will use that, and these small communities can still survive but would be more or less irrelevant.
Yeah, tried the same with Lemmy. A list of half done of dead communities, all of them with zero explanation of why and how.
Then i read about the codebase and people were struggling with SQL so performance is bad - i mean what? is that really the state things, where's the talent working on this?
I'm honestly surprised that both the design, performance and docs of these alternatives are this bad in 2023.
If this has to work you need an extremely simple and well designed landing page with a super simple explanation of how this works, the creation of a user should work on more or less all instances, and you should be able to easily meta search for the communities you want and get an easy overview quickly.
I think if the intent and structure of federation curves towards every household having its own server or something analogous the federated universe seems more sensible. But that does seem like a difficult curve until we have a universe where there's something like app-store level ease around self hosting your household services on something equivalent to a fancy home router is the norm. And that universe seems far off and actively not what the powers that be desire.
“ why would all problematic users congregate on the same server? ”
There’s a simple answer to why server blocking is a feature of mastodon. It was built to handle a real world problem. I.e there was early on a server that was a “free speech absolutist” server that server mods wanted to block at a server level.
I think many servers/instances post what they're all about and how they're different, and that should help. We're seeing all kinds of sites springing up to help navigate this, and I think in time it'll be sorted out.
That’s exactly the problem with replacing Reddit with a federated service. The opposite is true with replacing Twitter in which case it’s difficult to explain why a user has to pick any server at all.
If the goal is to talk about your interests with strangers (and not pull in your whole family and political and journalistic class), some minimal federation friction could be beneficial actually, as a soft filter. Maybe it's worth it to have some common pre-Eternal September spaces on the Internet, even it means not building unicorn empires.
The latter seems to be a mindset people aren't able to shake off for some reason, applicable to the given case or not. If a person can write and read two sorta coherent paragraphs of text for a internet forum, they are already on the level of being willing to learn and a little curious, above what the tech giants expect from their model user. There are enough of such people to have interesting and diverse discussions. Don't tell me Discord is easy to figure out. I'm not saying the devs should be working to make the experience easier, but it starts to be one of those topics where the discussion is meme-based.
You can tell search engine robots not to index your copies of different forums. I don't think they are viewable by default anyway? Avoiding small instances to avoid outages sounds like 4D strats to most users, who might prefer cosiness, personal connections and camaraderie, like on Minecraft servers, say.
The problem is discoverability. You don't know where to look for stuff or people. I wanted to check out this whole Lemmy thing and found almost completely dead (no comments) communities linked from their project page.
Where do I go? How do I search for something worthwhile? There's no central space to find it, compare communities etc. I was curious but now I am kind of put off.
It just needs to be posted as a main aggregator at https://join-lemmy.org, or even better, on the right hand side panel on every Lemmy instance.
I know this would be a centralized database of communities (which is probably frowned upon in the fediverse) but I don't think there's a sensible way around it.
Discoverability is a double edged sword. You'll find it, but so will every bot and scammer on the planet too. Then moderation will get ever stricter to get rid of the abusers and eventually the place becomes a mess for one reason or another.
I'm sure spam bots and scammers aren't waiting on indexes to appear and do their own, so providing an index can't make them more problematic they already are.
Trolls, bullies and harassers though will have better access to you.
I ran into this same issue. But I asked myself, if I passed through and left because no one was active, why not plant some seed content for the next person that passes through? I couldn't be the only one looking for smaller, more intimate communities.
I did this for a "dead" forum and after about a week of "squatting" there and posting for myself, other people started showing up and replying to it.
People should treat the digital silence of dead communities as an asset and free real estate to share their thoughts.
I checked and you can find a good number of communities by going to one of the larger Lemmy sites, clicking on the Communities tab and selecting All. If you want one directory from many instances (I assume this is what you meant by central space), someone would have to create it.
I wonder if the rise in smartphones as people’s default device has indeed reduced the likelihood of gathering such knowledgeable and worthwhile strangers together. Even many people with intellectual acumen and adventurousness, who might have tried an alternative community, are using devices that discourage longform text and logging into obscure websites.
Case in point, I am a member of some hobby and travel communities where, a decade ago, many people had blogs where they described their experiences with rich, detailed text. Now those same people hardly even touch a keyboard and are content to maybe just post a few pictures. They also are more likely these days to engage in political battle, since political polarization is so much greater than it used to be.
And they still state that it's easy and the other posters are scratching their heads as to why don't people understand this. "Whyyyy???"
They don't seem to understand that the average user doesn't find it easy to use. That's the biggest hint, but instead are acting like everyone is stupid but them.
That's why the fediverse is going to get a bump for a bit during this reddit situation, but then calm down as users aren't going to want to stick around. The same happened when others were leaving Twitter.
This fediverse situation reminds me of that Silicon Valley episode in which they're testing an app with a focus group. The group just doesn't understand the app and find it too complicated. Richard tries to explain the app to them, but it's too technical for them to comprehend.
> They don't seem to understand that the average user doesn't find it easy to use. That's the biggest hint, but instead are acting like everyone is stupid but them.
In all fairness those were the exact same comments when people moved away from Digg to reddit, only in the opposite direction. History is just repeating itself.
I remember vividly when the digg exodus happened, I don't ever remember a single person complaining that reddit, nor digg, were too technical/difficult to understand/use.
Both sites you simply clicked the register link, input your username/pass/email then done. The fediverse does not have a clean onboarding that lends itself to comprehension by a regular non-technical person. It seems way over the head of your average user. I've tried to get friends to check out Lemmy and it's a solid no-go.
If you need an entire matrix of understanding of a concept before feeling comfortable to use it then it is simply too much cognitive friction.
During the Digg migration, the main criticism I remember is that Reddit was ugly but still usable. For casual users of the fediverse, I think the criticism is that kbin, Lemmy, and Mastodon are actually too confusing to understand and therefore unusable, which is a key distinction.
Funnily enough, clicking the link currently results in a very long load time followed by a 5xx error. Federated social media has more than just the "understanding" problem, and the problems also compound on top of each other nicely.
Yes. The usual problems. Discovery, identity, storage.
Identity is the hardest problem. Yes, there's stuff like OpenID. But few non-programmers use it. We badly need some kind of ID system that doesn't depend
on a vendor that's in the advertising business.
Discovery is a problem mostly because Google has a monopoly in search.
The problem with federated storage is that storage is currently organized by "where", not by "what". There's the BitTorrent approach, where the address is the hash of the content.
There's IPFS, but it's 90% Make Money Fast with crypto and 10% actually storing stuff.
Whatever happened to the DOI concept?
There's distributed caching, like PeerTube, which works better. Every file has a home location, but if lots of people want something, it gets replicated. So if your cat video goes viral, your tiny server won't be overloaded. That solves the problem of popular content, but not the long tail.
And because the option to not index posts is enabled by default and a good chunk of the fediverse is opposed to indexing due to their irrational fear of bots:
I think the only reason people know about email supporting multiple providers is because its old, if email was made today, you bet its going to be centralized.
Heck people already scratch their heads when an email isn't hosted @gmail.com or @hotmail/outlook.com.
And the learning curve exists because unprofitable companies like Twitter and Reddit made centralization possible by burning money, the old internet was decentralized because centralization was impossible, it was too expensive.
If these failing companies die, federated services will be the mainstream again.
But as long as there is money to be burned to try to make a poor business model out of social interaction, federation is never going to be able to compete.
I've heard plenty of stories of people saying "my email is fred@fredsmith.com" and it being input by a customer rep as "fred@fredsmith.com@gmail.com". Or them asking "At gmail.com or hotmail?" and you have to explain "No, at fredsmith.com".
Except most normal people don't use their corporate email address for personal stuff.
I've always used a custom domain for my main email addresses and whenever I've had to give it to someone else verbally it inevitably leads to some amount of confusion. At the very least they get thrown off by the fact that I didn't say gmail/outlook/yahoo after the @ and I have to repeat the domain name.
> I mean that’s the whole point of federation right??
The point of federation is not to be dependent on oligopolies from a single country, one with poor grasp of and legal enforcement of personal privacy at that.
>It kinda loses its point.
Even if it comes to this, it doesn't. Another migration can happen at any point, but this time you wouldn't lose your social contacts - and in the future hopefully neither your content.
>But Google and Microsoft got in the game.
That's like your common anti-Linux argument.
>Search engine indexing for these federated services SUCKS.
Decentralized content/link aggregators just became popular so we'll see what'll happen, but personal social media shouldn't be as public for most people as it right now anyway.
> not to be dependent on oligopolies from a single country, one with poor grasp of and legal enforcement of personal privacy at that.
The oligopolies from a single country might get replaced by groups of instances who will claim they're holders of the Truth and who are the Service, while rest is dangerous. Nothing also stops countries from running own instances or compatible services by dedicated companies or people to either dominate or lure people out of the federation. A good campaign, attractive interface and features and people will follow.
The remaining option is to "run your own instance" but in time when the federation will most likely solidify enough it might be hard to do so. Hard as in getting people to join your new playground. What we're seeing right now it's a boom period.
Don't get me wrong, I do like concept of federated services but there's too many sides where it might fail in long-term and help create a highly tribalized echo-chambered communities
The problem with federation is that “HTTP browsers” simply just don’t work that way.
There is nothing you can do to make Firefox or Chrome to understand what a “federated network” even is. For a browser, an “Origin” is an extremely well defined concept. You can’t just decide to amend it because that’s how you feel it ought to be. If you can’t make mastodon.social be relates to social.mastodon in the browser, then you’ll have a rough edge. And that’s just to start. Federated services have so, so so so, many usability issues that all could be attributed to what a “browser” interrupts HTTP or the internet to be vs what is actually possible.
> Federated services have so, so so so, many usability issues
I like how you're talking about it like it's some abstract underground BBS style system that requires a modem, ISDN line installed, custom software running on a unix subsystem
I mean I found https://fosstodon.org/home signed up, clicked on local, liked posts, followed people, now I get a stream of posts to my home feed every day
Maybe your everyday facebook user might struggle with the idea of signing up to a site and interacting with it but I'm just not seeing why anyone who can sign up to discord or reddit would struggle with it
Try loading up an image on a different instance, or browsing that instance from a different instance. If you can figure it out, it's as slow as molasses.
That said when I see posts linked to Mastodon on HN I do notice that maybe the HN Hug of Death hits them hard as they are noticeably slower to load than my day to day interaction with Mastodon, maybe this is what you're seeing?
I'm sure it's fine if both servers have the resources but visiting a small instance from a different small instance and all the images load really slowly.
There's absolutely nothing stopping browser vendors from implementing ActivityPub. I don't think the browser is the right place, but browsers also implemented automatic RSS feed detection for more than a decade so it's not exactly unprecedented either.
Browsers don't need to care, anyway. There are some problems that need solving (notably, following/interacting with users on other platforms) but there are ways to do that. I'm not quite sure why nobody has come up with a good standard for a web-activitypub: URL handler, but once everyone agrees on a format adding support for cross-server stuff should be easy.
Web browsers can access webmail clients and mailing lists just fine. They even have mailto: as a means to cross the federation gap. All of these problems are implementation issues, not conceptual impossibilities.
I don't get these posts. I've yet to hear someone developing federated social media whose goal is mass uptake. Mastodon is designed to avoid viral posts for instance. I use latex and vim to produce pdf documents. I wouldn't write a long post saying "I don't see this replacing ms office any time soon".
> I don't get these posts. I've yet to hear someone developing federated social media whose goal is mass uptake.
In which case it's being mis-sold. It will never replace the thing that people are claiming it will replace.
> Mastodon is designed to avoid viral posts
"viral posts" aren't inherently a bad thing. It's part and parcel of some of the good aspects of social media - the sense of having a collective conversation among a large group. Fragementation is fine up to a point (We're already fragmented between half a dozen large social networks) but I want people to find me easily if they are interested in what I'm posting.
And why do you want to be found easily, really? In case you get your 15 minutes? I think the goal is to build community. Large numbers and reach are actually not so important.
If you're thinking of federation in terms of "apps" yes it's doomed.
Like with email, you have to think about it in terms of "service".
The "gravitate to one provider" problem is there (as with gmail), but those who care can still have their own email services or use reasonable third parties.
The tragedy of course is that very few people care.
Well since circa mid-2010s the Internet is apps. Or app-like centralized web sites (like Hacker News).
People want to install a single “Lemmy” app on their phone that shows them all Lemmy content and lets them post to all instances, using a single account and a *simple* user interface. If Lemmy cannot provide that user experience (abstracting away the federation), then yes it is doomed.
I have been actively looking for alternatives since the Reddit API fiasco started but I just don't have time to sit down and grok Lemmy or any of the other federated instances. And I'd argue I'm more tech curious than most. If this is such an obstacle for me, I imagine very many people won't bother either.
If you think bittorrent like protocols are better suited for your usage, there are networks that use that: nostr and scuttlebut. They are not part of what people call the fediverse, because the underlying data format is different, but you can still try it.
To answer the question, and I have just a superficial understanding of the concept, the main issue with decentralized vs. federated protocols is that you have to store the whole (or quite a large chunk of the) social tree locally on your device. That's a lot of data that you don't really care about. It might be fine if you access the network from a computer with plenty of space, but a mobile device might be more problematic.
Federation doesn't mean anything like that per se, you're talking detail of a particular implementation (I assume Mastodon but what do I know).
Federation just means multiple independent servers/domains can talk to each other, such as with email or Matrix, without the need for central authority (other than for discovering each other).
The concept that you get wrong is that nobody mirrors content. Servers receive the content from the one of the original content creator, if they exist in its recipients list. Public content usually is distributed to all servers that the original server knows about. However it's always a PUSH of the content.
I think it's worth pointing out here that we (on HN) are not average users.
I have had a number of actually average users asking me about the Reddit blackouts recently. Most of them don't know of Reddit beyond the first party app. 3 or 4 didn't even know Reddit had a website. None of them could even comprehend how you could use a third party app to access Reddit and their minds were blown when I told them that Reddit didn't have an app at all originally and the Reddit app today was originally Alien Blue.
For most people, the word "service" doesn't even enter into their minds. It's an app. It either does what they want or it doesn't. The inability for many in our group, including myself, to comprehend and understand this is why the fediverse, Linux, crypto for non-investment/gambling purposes, etc. continue to fail to gain much popularity. For most people, they don't care how it works... they only care that "it just works".
One problem I see is the lack of cohesiveness between clients. What one client/app chooses to name a “repost” is different than another client. Icons for said features also vastly differ. So it’s hard to build unity through the ecosystem without a common language (visual and vocabulary). This is just my brand developer’s take.
This is a big part of why I'm skeptical when folks point to email as an example of a successful federated system: most of the common language and conceptual model was established thousands of years ago.
I believe it's safe to say that in a thousand years ActivityPub will be equally well understood and uniform. You're just here for the ride. Try to enjoy it.
A big problem is that you don't actually see content on other instances. Also, you can't link accounts between instances (EDIT: you can post and subscribe to communities on other instances, but it's not obvious. You can't link 2 accounts on separate instances though)
This is what I initially thought the Fediverse was like. Right now it honestly just seems like many people running completely different servers and communities using the same software.
(This could even help with the decentralization aspect, because ideally the content and accounts would be replicated. So if one instance goes down, its content is still on connected instances, and possibily "ghost accounts" from that instance which let users there continue to post).
I think grouping different communities together would be a great feature, but you'd still have to find someone to curate and collect those groups, and host them somewhere.
"Why are all of these communities there" is a bit of a weird question, in my opinion. Why is there a news.ycombinator.com when Reddit already existed? Why was reddit founded when Digg was around? Why aren't don't Outlook and Gmail combine forces into one? Why would anyone want to go to stackoverflow when expertsexchange and hundreds of forums already existed?
Every server has its own rules, that's why. Some are more tolerant of abusive behaviour under the banner of free speech, some will use bans liberally to maintain their community. Some servers ban NSFW posts, others don't care. Some servers block some other servers for abuse, others don't. Some communities are fine with meme posts mixed into serious posts, others want to stay on topic.
Even on a centralised platform such as Reddit there have always been duplicate subreddits because of people who disagreed with another subreddit's rules or moderation practices.
Nobody owns rust, except as it may exist on their server. This is basically like asking why (e.g.) rust, and rustcode, and rustnews are all different subreddits and why don't the same links appear in all of them. Because they're different. If you want to see all of them, you subscribe to all of them.
Why should 1 person or instance own rust? Why should they be linked. They aren’t the same thing. The namespaces are rust@programming.dev and rust@lemmy.ml so on obviously they aren’t the same thing.
You’re confusing instances with communities. Instances are servers. People create and mod communities. They shouldn’t be linked because they have different mods and rules.
> A big problem is that you don't actually see content on other instances
What do you mean? On lemmy you simply click "all" and see a feed from all federated sources, on kbin that seems to be the default. When you open threads/posts you see comments from all sources. The other stuff doesn't exist yet as the platform is very new. You can help.
this isn't strictly true and the weird edge cases add more mental energy than conforming to the expectation of "all of the content I was expecting".
For example, clicking on one of the profiles I follow, and then their followers. This will only show the profiles I also follow on that remote server. Do that again, and it gets even worse.
Where instead, the "thing you expect to happen" is to list, 1:1, their follow list as it exists on the remote server.
Other things, for example, not fetching someone's post history earlier than before you followed them, when you're browsing their profile. e.g., a user with 100 posts that you just followed.
I'm aware of the technical reasons (you weren't subscribed), and the open source reasons (not enough volunteers to fix the bug); but both are inadequate to a new user, frustrated and lonely, struggling with a lack of network effects, where having a delightful app might save that experience.
> This could even help with the decentralization aspect, because ideally the content and accounts would be replicated. So if one instance goes down, its content is still on connected instances, and possibily "ghost accounts" from that instance which let users there continue to post
Or, IRC. For all the major support systems (most I think now have moved to discord), they all had a freenode channel, but what if they'd all chosen different networks, so you'd need to connect to maybe a different network for mysql, another for postgres, another for php, another for rust, etc...
To make it more confusing though, all of those networks EACH have their own smaller channel for those technologies, just the official one is on x server.
Reddit is awesome because it's one place to get ALL content related to a niche under one umbrella, and because you can use awesome 3rd party apps -- the next reddit replacement doesn't need to be federated, but it could be community-ran as a sort of DAO, where users can vote on features/etc...
It also costs money to host instances. And I wonder how will we be able to monetize them. Will we have ad-supported and premium instances? Is that possible with Lemmy?
As much as people don’t like websites making their behavioural profiles and selling that data or advertising with it, most people are also not willing to pay for access to social media. There are many who would be willing, but will there be “critical-mass” many?
So it will be very interesting to see how the instances can be made self-sustainable. Right now (but correct me if I’m wrong), I think it doesn’t make financial sense to run one. People can do this for a while to support “the movement”, but I wonder what happens when “the movement” is over.
> Here’s the biggest kicker for me. Search engine indexing for these federated services SUCKS.
That's a rather odd argument as "the biggest kicker". I'm fairly sure the teams at Google and other search providers are very aware of the fediverse and its rising relevance. It might be hard to find something today, but eventually the search engines will be improved so that they understand federated content.
Honestly these all seem like solvable albeit hard problems.
Learning curve is a ux problem. Yes federated is a bit harder, but i think this is more the general problem that open source is terrible at ux, especially when it doesn't have sonething to copy, not a fundamental federated issue.
(Cross-instance) Authentication - definitely hard, but seems solvable as well. Whether with fancy crypto or maybe some federated oauth thing.
SEO - meh, who cares. Do people really use these sites for googling. Also seems like could be solved by getting everyone to shove some rel=canonical all over the place.
Personally i think the fundamental issue (as opposed to just engineering challenge issues) is that its hard to modify/adapt federated protocols. I very much agree with moxie's critique of federation at https://signal.org/blog/the-ecosystem-is-moving/ . Everything else is a simple matter of programming - an inability to rapidly iterate is on the other hand is an unsolvable problem.
"Federated App" sounds odd: protocols are federated, and apparently they are criticized in the article, not software.
The "Learning curve" argument suggests that oligopolies tend to form, so it is not worth bothering. But that applies to pretty much any organizational activity, not just federated protocols, and often people try to push back against oligopolies or monopolies; often it is desirable and possible.
The "Authentication" one only applies to web-based clients, but you can login to remote instances even with those: converse.js supports it for XMPP, for instance. Possibly not in the way the author imagines it should though; perhaps that is closer to OpenID/WebID/OAuth. Still, generally users are not meant to login to every instance directly: even with just Google and Microsoft, you do not login to a Microsoft server to send them a message from your Gmail account, as you do not login into the receiver's email account to place a message for them. Maybe this section should have been titled "learning curve" instead.
The "Search Engine Optimization" probably describes the current state of things (I am not familiar with that), though does not apply to many federated protocols (e.g., I do not want my email or XMPP messages to be SEO'd), and I do not see a technical reason for that when it may be useful (e.g., public blogging or discussions); certainly does not sound like a drawback of federation itself to me, but maybe that of certain search engines' policies.
I personally usually saw discoverability/exploration as the primary issue with both federated and distributed protocols, especially when it comes to social ones, and hoped that RDF with FOAF will help with it. But ActivityPub and Mastodon seem to handle it rather well now, out of fairly widely used protocols.
As far as email is concerned, one reason why you don't see as much as self-hosting or any frankly, is due to deliverability. Doesn't help to have your own mail server when all the mail coming from it is classified as spam and filtered out of inboxes leaving you to have to use a "trusted" hosted provider to ensure your email actually works as you'd expect.
Is it really a problem or cargo culting? I certainly thought it was a problem, but I know a sysadmin that runs two small businesses's email on prem and he says that deliverability, even to major providers, has never been a problem, as long as you set up everything right.
I don't know how cargo culting figures into this. Nobody is setting up cardboard boxes and thinking it's a mail server :).
What does it mean to set everything up right? Out of the box postfix, sendmail, dovecat, etc., no longer generally work, even with DKIMs, and other necessary domain verifications.
I have heard of some admins opening cases with some of the large hosts to ensure their mail goes through, but seeing how those are all big cos with little no guaranteed lines of contact or follow through I wouldn't hold my breath on it working or being a long-term solution as they can easily blacklist you at any time with you having no recourse.
I'd love to hear a counter argument, as it's also the same thing for basic transactional mail sending, there's a reason everyone switched to SMTP hosts such as Mandrill, Sendgrid, and AWS, deliverability.
The article sums up my current frustrations with federated platforms. Maybe it just needs some time for reliable communities to be known. But until then, I'll be happy to use centralized platforms. Alternatively, people can adapt the culture of self-hosting their own instances. Then you don't have to worry about reliability and content moderation.
The Federated App Problem has already been solved a couple of times using replicated gossip networks of signed messages. Sometimes they use append-only logs, other times they use only relays, and the implementations are all highly creative and experimental.
The trouble with decentralized federation social media servers is the weight of the network falls upon whoever volunteered to host the program at the domain name in the first place, and sometimes the weight is too much to handle.
The solution is, as we all already know, self-authenticating messages that do not rely on any one server. Do we want to call this Scuttlebot? Do we want to call this Nostr? Does Bluesky even qualify? Well, I'm sure we'll figure it out eventually with enough minds trained upon the problem.
I thought this would address the scaling of federating activitypub messages, but it's really just problems from an end user perspective. Those problems we can easily solve. Mastodon auth is already a thing, allowing the same app to authenticate against multiple types of backend software, and oauth can also be enabled to use a central identity.
When I hosted a fedi node I disabled search engine crawling with robots.txt, but it can easily be enabled.
Learning curve is all about giving it time and earning experience in accessibility.
But regarding the scaling out of messages; I am writing a small activitypub app in Python, because it's just so fast to prototype in, but I want the option to rewrite heavily used components in another language. So instead of passing pickled or serialized Python/Ruby code around background job processing, I aim to only pass ActivityStream objects, simple JSON that is language agnostic.
Just put it in different queues depending on its destination. Then any background job processing program can process the queues and put the data into DB, or whatever needs to happen.
The handful of things mentioned in that article are definitely frictional, but I think people are willing and open to learning if they start to see value.
Smaller Lemmy instances feel more personal to me than subreddits did, and I think that is something people will find valuable.
The "federation is unintuitive" is not an absolute problem. URLs are unintuitive. ".com"? What on earth is that? All new systems require you to learn. If the content is there people will use it.
People cluster around one or a handful of servers... that doesn't make federation pointless. People keep thinking of federation as an absolute, everything must be evenly distributed, but it isn't. Email is the perfect example. Most people use gmail or outlook. Imagine if outlook stopped sending or receiving from anyone else. The point of federation isn't to ensure even distribution of content, it is that it can and must be accessible from outside. It cannot be siloed without becoming useless. Even if 90% of users use one or two servers, the minute that server blocks outside access that server becomes a pariah.
Authentication... Why would one server let you login to an account on another server? I see the problem from a user perspective, but this is simply a misunderstanding of the architecture. In any case, hubzilla (zot protocol) has an interesting solution to this, and Nostr, which is not a federation, doesnt have this problem at all. I think both of those protocols are fantastic.
SEO, with only one engine mentioned... That sounds like a google problem. This technology is cool, if the content is there and google doesnt deliver it, it will affect them in the long run worse than the federated system.
I personally don't like the way activitypub works, or even like federation. I don't think community creation, maintenance and engagement are very good use cases for federation. I don't see why a server in a federation needs more than one community, as is the case with Lemmy, and I don't think communities, let's say HN for example, benefit from federation, because it encourages drive by engagement from people with no real commitment to the community. Reddit has this problem, spillover engagement, and the subreddit where it is not a problem aren't really communities, theyre just tags. "Memes" isn't a community, "chicagobears" is. The former could just be a tag in a federated system, the latter doesn't benefit from being in one. At best, two autonomous communities that have overlap could benefit from it, so having software that supports it is good, but on by default federation is not a good paradigm for forum like community software.
The big one here that is compelling for me is Identity (sort of Auth as touched on in the article). My understanding is that there are multiple working groups trying to tackle this, but I don't know of any federated networks with a sizeable userbase that has implemented any of the decentralized identity mechanisms yet. (Please let me know if you're aware of one).
That said, Matrix is quite large, and growing, and being adopted by multiple national governments. And I know they want to eventually tackle decentralized identity, there are open issues and some brainstorming about it.
I am hoping that a federated alternative to HN will emerge because this is the only non-federated social platform I (will) participate on, post-reddit. And probably the only one that has reasonably stood the test of time in the face of tech capitalism, and it's certainly unique in that regard. I like building software that is aspirational and going to stand the test of time, and I wish I had spent more energy til now working to the same end in terms of "knowledge-bases" ala those that have grown in now-private subreddits and semi-private Discord servers.
That’s interesting. I don’t hear some of the other issues, such as updating and propagating changes to servers and clients. (Although, I wonder to what extent AI can help with that).
Prisoners step out in the sun to find themselves confused in a never ending maze full of branches in all directions. They have difficulties finding shelter and food.
They walk back to their well understood confines, in relief. The door doesn't lock, but they choose to stay either way.
Assuming you are not a prison operator this smacks of complacency.
In the digital age where rules are there to be broken fast, any and all complacency will be used against you (us).
The era of web 2.0 social media, damaging as it was, is drawing towards a close.
The next phase will be an ever tighter "embrace", with airtight control of digital identity, access to financial transactions and unprecedented AI manipulation. Of-course the inmates will enjoy Cloud Atlas type amenities.
Freedom was never easy and is not an exportable good either. It is either in your real constitution or not. I just hope that there is a critical mass of countries that actually mean it.
Mastodon (and Lemmy) had the opportunity of its lifetime and failed to capture it. It honestly feels like crypto with lots of smart people wasting time on a doomed concept.
I feel like we are comparing apple and oranges here. Mastodon and Lemmy are free project, run by a community and, in the case of Mastodon, backed by a non-profit. Depending on who you ask, they don't really give a damn about capturing the opportunity. It is not a for profit venture where capturing new user is essential and critical. It is about creating something they think are cool, neat and want to use.
I do see the potential advantages of federation to prevent the same fate that overcame Twitter and Reddit, but I can’t figure out how to make a user’s server choice actually meaningful. Without solving that piece of the puzzle, I don’t see any practical advantage to federated services.