So let’s take this actual example: There’s !canada@lemmy.ml and !canada@lemmy.ca. They talk about the same thing, but are treated by the current federation implementation as separate instances.
How would you feel if there was a moderation feature to import another federated instance’s community into your own, so that the posts from the other instance automatically show up in the same feed? That way, you only have to subscribe to one community on one instance, but you get content from multiple instances. I’m not talking about crossposting or mirroring/duplicating posts between communities, only displaying the posts from another instance the community’s home server federates with, with moderator discretion.
Interesting, I hadn’t contemplated the idea of topic-dedicated servers. So looking at multiple servers through lemmy.ml is functionally equivalent to using a local client to pull the data from each server separately. I don’t think that’s a bad idea, it’s just not what I’d assumed.
As I said in my initial post, clearly I have some things to learn! I’m not a programmer, just a superuser with an interest in finding an alternative to Reddit’s increasingly invasive and restrictive platform.
That said, I do still like the idea of giving subs the option to “merge” if so desired. It’s an interesting layer of complexity to add on top of what I’m realizing is already a fairly complex system.
Small nit: it’s not like pulling data from each server separately because you can have one user account on one server and vote / comment / post in communities on all the servers using that one same identity.
So what’s the role of the server in account creation/maintenance? Others could take the name “Flelk” on other servers and parade around pretending to be me, meaning readers have to check the domain name to confirm whether it really is me? If my “home” server deletes my account, does that erase all my votes, comments, and posts on just my “home” server, or on all servers?
(I know you folks are busy actually building this thing, so if there are better ways to get answers to these questions other than asking you admins directly, please feel free to point me in that direction.)
The domain name pretty much is a part of your username, just like it is a part of an email address. Impersonation is unlikely as your name is displayed as @Flelk@lemmy.ml whenever your comments or posts are viewed from another instance. Your bigger concern for impersonation is somebody using lower case L instead of uppercase i and stuff like that (I and l, can you see the difference?)
(Unless you choose to have a display name, that’s for some reason the same on all instances. I assume even on the same instance there can be multiple users with the same display name. Not sure why people use it tho, it could easily get confusing)
This is how your account looks from lemmy.ca for example.
As far as I’m aware, if you delete your account it only gets deleted from your home instance. A removal request gets sent to other instances, they could in theory refuse it. I’m not sure what data the other instances store and for how long tho. I think your public stuff like comments and posts etc gets stored?
Now that’s interesting, only my posts in !lemmy@lemmy.ml are visible on lemmy.ca. My first post to !reddit@lemmy.ml and my post in !startrekmemes@lemmy.ml don’t show up at all. Do those other communities not sync to lemmy.ca? What’s going on there?
Your own server is the one through which you interact with all others. You just talk to it, and then it talks to the other servers.
Yup, the domain name is part of what defines your identity. I would expect that eventually we’ll have more interface options to ensure it’s not too confusing who’s who (especially since there’s your real username and then you can also set a display name) but it’s one of those things that isn’t really a problem until it’s a problem.
Deletions in the fediverse have been a big deal in past. The tl;dr is that your “home” server would send out a “hey delete this” notification to all the other servers. By default they will of course do that, but you can see that it’s conceivable that someone could make a malicious version of server software that wouldn’t.
I am not a dev on the project so I am happy to pitch in answering Qs. :)