or maybe some other terminology would be better? lots of people get confused when you ask them to choose an instance, sometimes I think even the word “proxy”, “host”, or “hub” is simpler
the specific terms aren’t my point, just a discussion to see if we can come up with a better name
I don’t think it’s about the term, “server” and “instance” both make sense to me. The issue is that the fediverse itself is pretty confusing.
The basics? Great: it’s vaguely “IRC but persistent”, all good.
But for starters it’s hard to keep track of which instances actually exist - new ones pop up and old ones die at the drop of a hat.
Then there’s differences in feature sets (lemmy vs kbin and whatever else) that happen to be ActivityPub compliant or whatever. kbin notably doesn’t federate downvotes, for example. And all this software is still relatively immature.
Then there’s the actual “who federates/defederates whom and why” debacle. This results in a lot of obvious and some less obvious visibility issues.
Then there’s (other) individual instance politics.
Then there’s the “meta” about all of this, which is getting confusing.
A couple of these will have parallels on e.g. Reddit - I assume this is the natural comparison to make and will keep being so for a while - like sub drama and the relationship between subs. But because the FV has this at the instance level, (and each instance has many “subs”,) it’s a whole level up in complexity.
Then there’s how all of this makes for a pretty un-reddit-like experience - and Reddit is not the king of polish, either. While Reddit has duplicate subs, it doesn’t have a design that almost automatically causes them to be created and distributed, across instances without actually correlating them afterwards. The end result is that subbing or blocking any one community will likely involve doing that manually on several instances, which is stupidly inconvenient. Also discoverability is much trickier which is worsened by the low activity.
My point is: call it what you want, but a) I don’t think that’s where the confusion is coming from - that’s just the fediverse being confusing (and outright clunky in many regards), and b) obligatory XKCD “Standards”.
Personally, I’ve been using the words “site” or “website”, because I think highlighting the fact that each instance is its own independent website clarifies the issue to a large degree.
But you’re 100% right. It just doesn’t alleviate the sense of overwhelm people feel. And I don’t know that anything really will, except for repeated and continued exposure, because networks of quasi-independent actors are complicated things, and the world is now full of people who have experienced the internet as little more than 5 insulated websites. The mental model that people have for social media is just “everyone’s reliably using the same website as me”. The idea that different social media websites are communicating with each other, and also that those social media websites don’t have a billion accounts – and don’t need a billion accounts in order to be viable – is just… alien. To the point where even those of us who are engaging in the experiment kind of sweep the essence of the space under the rug, you know? Everyone treats “Mastodon” as a singular location. This here is “Lemmy”. “kbin” is over there, at a particular URL. If we treated the rest of the internet with this level of abstraction, I’d have to tell you that I was “On Firefox” right now, or telling my wife about this meme I saw “On macOS”, or “at my desk”.
And like, sure, some of us have a deeper internal understanding of federated social media. We heavily used IRC in the past, or get grok how email works, or whatever, but the fact that we still all kind of collectively brush aside the heterogeneous and quasi-independent nature of the network when actually using it in practice I think speaks to just how heady it all really is. And I’m not sure there’s a linguistic solution to it. It’s just an incredibly messy space in a world where people crave simplicity.
But that goes against the original point of the fediverse IMO, which was to make a resilient social media platform where it doesn’t really matter what instance you join, you’ll get the same content. If we treat them as separate sites, then we should probably remove the federation entirely and just have duplicates of communities at each instance and just handle things on the frontend with links.
So I think it failed at its original goal, and now it’s some weird mix of separate sites and a large, decentralized ecosystem. People aren’t sure if there should be separate, smaller communities or larger communities organized by instance, and we end up with a weird mix of the two (multiple, large-ish communities targeting a similar goal).
I’m not here because I think the model is the right direction, I’m here because it’s better (for my priorities) than available alternatives. What I want is decentralized Reddit (i.e. one namespace for all communities, but not hosted in one location), but my options are centralized or federated services. I want the complexity abstracted from me, not in my face like it is here on Lemmy.
If that was truly the original point of the Fediverse, it failed at the design phase. The way content is hosted and passed around has meant it was always going to be a constellation of independent nodes, each doing their own things. There’s nothing in the fundamental design of how these networks work that points to them being a networked simulation of centralized social media. And the repeated attempts to make it work, or at least look like it works, that way has resulted in exactly what should expected from trying to jam that square peg into this round hole: A poor and messy simulacra of centralized social media.
It has always been – and this is necessary by design – a weekly interconnected network of social media and networking sites. That’s the true, fundamental nature of the space, based on the engine powering it. Trying to pretend otherwise is just adding complexity on top of it, not removing it.
Maybe I’m misremembering the original “marketing” about it, but at least Lemmy has this to say:
That’s a lie, you can only access content that is federated with you, and there’s a complex set of relationships between instances where you will always be missing some portion of the fediverse (i.e. if C blocks A, and C posts to B, users from A don’t see that content on B, but users from B do).
So I’m not sure if it was siloed by design, but Lemmy was designed to replace Reddit, so presumably the same notion of what Reddit means (people congregate into communities, instead of instances) is implied:
But I obviously can’t say for certain whether the original intention was to make tons of Reddit alternatives that all kind of connect to eachother, or to make a centralized Reddit alternative that is decentralized to prevent any one node disappearing from wrecking the network. If the former, I don’t really understand the point, and if the latter, I think it’s the wrong architecture.
Regardless, it’s better than Reddit, so I stick around. I assume the same is true for Mastodon and Facebook.
I mean, I wasn’t here a decadeo ago or so when the groundwork of the Fediverse was being laid, so I don’t know how it was originally “marketed”, but people make things without understanding the true implications of their decisions all of the time. And the current crop of leading products in the fediverse are a generation or three removed from the original designers.
People build on top of stuff with goals that are off-target of the original goals of tech. Building a bunch of square pegs and ramming them through round holes just, ultimately, results in those pegs either not slipping through, or having their corners cut off.
Yup, that’s the way FOSS goes. Build what interests you, and make mistakes along the way.
Eventually the community seems to arrive at a decent solution though.
I’m really interested in working on a project that makes a proper decentralized Reddit/Twitter.
Have you heard of Lime Reader?
I don’t really know the specifics but it’s a different approach to decentralized reddit but it kind of came too late to get exodus traction.
Do you know how it works under the hood? I didn’t see an obvious git repo in the few minutes I looked around.
I actually just started hacking on my own because I noticed a library I want to use (Iroh) finally has a release with decent documentation and a relevant example. The main idea is that each app install would help host the data, support searches, etc, so there’s no single point of failure, or any real requirement for people to host larger instances (maybe just some relay hosting).
I’m guessing there are several floating around, but I figured I might as well try my own to at least get familiar with the library.