Can someone explain to me how it makes any sense for Pixelfed to be a thing that has its own accounts, rather than being an ActivityPub client whose GUI just happens to be image-focused?
@jwz It's a distinct failing of the system, I think. In theory, you can follow mastodon accounts on pixelfed, and vice versa, but the affordances live with the base service/UI that you sign up for. There's no federated identity where (e.g.) a mastodon account can be used to give you combined access to a pixelfed (or other) service.
It's one of the reasons I haven't signed up for any other Fedi services -- I can't be arsed to manage multiple logins.
Mastodon, Pixelfed, Friendica are 3 examples (of many) of completely different software platforms that serve completely different intents and have different features sets that mirror/mimic their contemporaries like Twitter, Instagram, and Facebook. The difference is that the former share a common backend component that allows you to interact with each other across the various platforms, whereas the latter do not.
/Continued
You don't sign up on all of them. You sign up on the one, or few, that provide the services and features you're interested in using. And you use that, or those, to interact with anyone on any activitypub based server, running any of the dozens and dozens of different activitypub based software platforms.
That's not a flaw of activitypub. That's its primary feature.
/Continued
If you enjoy a Twitter like experience and want that kind of an experience on the fediverse, you sign up for a mastodon account or one of the other Twitter style micro blogging activitypub platforms.
If you want more of a picture focused Instagram style experience, sign up for a pixelfed account.
The beauty of the fediverse is that it's really whatever you want it to be. You got options. Which can be overwhelming, for sure. I think it's great though.
or, hear me out: you sign up for one account, like email. you can then use whatever apps you want with that same account.
mastodon, pixelfed, etc don't all need to be separate accounts on separate servers. just like we don't need a separate Web for chrome vs firefox -- they let you browse the same Web.
I feel like I've had a reasonably decent grasp of the way things currently work in practice. I'm trying to make sure I'm wrapping my head around how it would work based on this new information you've given me.
Would you liken this to like a single sign-on system? When someone follows you, they follow your sign-on handle and they would receive anything you posted from any service while logged in with that handle? Wouldn't that need a centralized sign-on service of some sort?
@finner yeah, that's pretty much it
you have an actor on social.example and you use clients like mastodon.example or pixelfed.example as you feel like it
people follow your actor on social.example and your actor on social.example can send out arbitrary activities to those followers. these activities are generated by the clients (mastodon.example, pixelfed.example, etc) and handed off to your outbox on social.example
this is like using thunderbird or outlook which don't do raw SMTP themselves.
ActivityPub is a lot like RSS, where you follow people and their posts get sent to you, and you view it in your Feed Reader (or in this case, fediverse app).
This is different than a federated single sign on system, such as OpenWebAuth or OpenID. There are some fediverse platforms, such as Hubzilla, Streams, and Forte, that support federated single sign on. In that case, you can use your existing fediverse identity to log into other websites, and view and comment locally.
@scott yeah I've read a bit about the Zot protocol with nomadic identity for hubzilla. It's pretty neat.
I think the one reason that got developers so hung up on the application server model is because everyone assumed that the client needed to have an inbox URL to listen to requests all the time, which led to the idea that AP *requires* the push-based flow.
This will hopefully change this year, though. I started working on the ideas I've laid out on https://raphael.lullis.net/a-plan-for-social-media-less-fedi-more-webby/ , I just hope I don't go broke before I get something usable out of it.
I was just talking about this here. I think I even mentioned.
https://fedia.io/m/fediverse@lemmy.world/t/1437745
Is there any future where we could see this being the future of activitypub?
@TheFederatedPipe it could happen in maybe 10 years, idk — it requires philosophical and cultural changes on parts of fedi devs, to stop building monoliths that clone popular centralized apps, and start building for the Web instead. things like facebook and twitter are not built with an open decentralized web in mind. they are built for keeping you in-app. copying their design means copying their assumptions.
@TheFederatedPipe the other thing that keeps most devs from adopting C2S currently is a lack of existing infrastructure, and some unfilled gaps in the user experience that require extensions to fill in. we also need wider standardization of things like authorization and authentication, especially cross-domain. having a reference server would probably help a lot, but it’s a chicken-and-egg situation a bit…
You know? As much as I love the fediverse, things like this kind of disappoint me? I'm not sure if that is the correct way of saying it. Like not everything is actually how it was promised. But I guess is better than the other option, hopefully we get a lot of this things sort out.
@TheFederatedPipe yeah, it’s the idea that things could be better but they arent :/
twitter imo set back online social communication by at least 15 years
Agree, I kind of feel the same about Mastodon sometimes, like they in someway they make the development for a brighter fediverse harder because they are not willing to implement things making other services not implement them out of fear of not working with Mastodon.
@TheFederatedPipe @jwz @darkling @finner @trwnh We need to have more implementations of the ActivityPub API, and in particular the read-write parts of that API.