Starting your own instance doesn’t solve the problem of big communities being reliant on the one specific instance they are hosted on to not go down or rogue.

I imagine it like friend requests between communities: x@instance.a, all-about-x@instance.b and x-is-great@instance.c could send each other friend requests and merge into one federated meta-community about x. Then if one instance goes down the other two are still there to keep the meta-community alive, and if one goes rogue the others can just unfriend and keep going without it.

The nice thing about manual federation is that the communities don’t have to have exactly the same name, and the mods can keep malicious or troll communities out. And ofc you could still have client-side control if you want to, e.g. add or remove a community just for you locally, or create your own local meta-community.

The database, storage and network are usually the bottlenecks in these kinds of websites, not the programming language. It might add a few ms of latency, but the big lags come from congestion or bad db queries.

What comment?

Wow that looks good, thank you very much!

Looks good so far, but two major features I want from a note taking app are still missing: Handwriting and table calculations. If they can add good support for those then I’ll definitely switch!

Afaik it is all connected to the domain name, so they could definitely start to impersonate any .ml instance. Other instances could detect that the signing key for federation messages changed, but that’s about it. Their admins would probably have to block/defederate them manually.

Not really. Most centralized services are accessible via multiple domains, e.g. for different countries. This would just disable one of them, but users could still use another to log into their accounts. For the Fediverse it “disables” an entire instance, cuts it off from federation and locks out users.

Lets not put a positive spin on a situation that exposes a weakness of the current system. The federation protocol needs to be able to handle these things gracefully, like propagating domain changes and migrating accounts between instances!

A domain takedown was never able to shut a server down, not even with centralized servers. Most big services are accessible via multiple domains of different countries, and this would just disable one of them. But for the Fediverse that means that they also “disabled” an entire instance with all its users.

This actually shows us that relying on domains can be a problem for the Fediverse! Imo we need to upgrade the federation protocol to be able to handle these things, like propagating a domain change or migrating accounts to other instances.

Using p2p for messaging is really nice for decentralization, but it has the major downside that both communication partners have to be online at the same time to find each other and transmit a message. So you might have to wait for it until both look at their phones at the exact same time. On top there are privacy issues, like being able to see the devices and public IP addresses of other users.

Imo its just not practical and robust enough to be used by millions of non-techy people.

Yes it’s also federated, so you can access their instances and communities pretty much like the Lemmy ones.

Fedirect

This is what I hope happens to Twitter & Meta.

Your right to choose is the same as everybody else’s right to choose. You can decide to post something, and others can decide they don’t want to see it. Decentralized just means there is no one entity to make those decisions for you.

I doubt it. While the server to server communication is standardized with ActivityPub, the server to client (app) communication isn’t. This means different Fediverse apps have different APIs to login, fetch and create posts and so on. It’s unlikely that we’ll ever have a standard API here, because the apps work diffently and have different needs for their APIs. Maybe some kind of common denominator, but that probably wouldn’t be feature complete on most apps.

So the one app to rule them all would have to created custom API Implementations and UIs for every Fediverse app, which would be a loooot of work and probably a worse experience than specialized apps.

My server is a DiskStation, so I use HyperBackup to do an encrypted backup of the important data to their Synology C2 service every night.

The thing is, people might not know that they have to look for the Docker setup. Now if they search for “Lemmy on Synology NAS” they land on this guide, telling them to use Docker.

Same. Having all their custom software available and just one click away is amazing, and with Docker you can install everything else just like a regular server. It’s the best of both worlds imo.

If you don’t have any restrictions (limited subdomains, service only works on the server root etc.) then it’s really just a personal preference. I usually try paths first, and switch to subdomains if that doesn’t work.