this whole discourse is why ive been yelling 'federation happens in the client' for the past 2 years btw
clients are the aggregation point and the interface to the network. thats why thats the point is where censorship can be meaningfully applied, and thus the point of contestation
people say the fediverse is more decentralised than bluesky, and thats true... but people saying that are usually thinking about fediverse in terms of servers. the fediverse is also much more decentralised in terms of client usage, and i think that actually matters much more in practice than the decentralisation of servers
@laurenshof It’s such a bummer to me that most clients use the Mastodon API, not an actual standard. I’ve blogged about this and brought it up at FediForum, but it doesn’t seem to be on most people’s radar as a priority.
yeah, it's a classic embrace-and-extend maneuver that's really slowed innovation in fedi (as well as reinforcingMastodon's dominant position despite their lack of progress). Hopefully priorities will change!
@laurenshof if you mean that in the sense of how many parties need to acquiesce to censor/deanonymize the whole network, then yeah, sure. But the client is a server, and that's also true of bsky. The technical difference is pretty marginal in that regard. Bsky is centralized in terms of power and relationships, which is the only way that has ever genuinely mattered.
@nikclayton Interesting, thanks for sharing the link. I don’t think it needs to use GraphQL, but I’m glad to see the discussion!
@thenexusofprivacy Yep, I don’t think Mastodon intended for this to happen, but it’s the natural consequence of private APIs.
@mackuba @quillmatiq @jdp23 @laurenshof that's the actual reason why twitter closed its API. Tweetdeek had so much marketshare that, at one point, the could decide to change the server. They owned the user base.
@manton I think the GraphQL bit is a red herring.
Efforts to create a successful alternative to the Mastodon API for user agents to use are doomed to fail unless Mastodon GmbH adjust their priorities and fully commit to implementing that API.
Otherwise you're asking client developers to support a new API that will only be usable by a tiny fraction of their user base.
@nikclayton It’s difficult for sure. It would take at least one popular client supporting both the Mastodon API and another standard API, probably for years while the API gets traction. That’s what’s nice about using Micropub, though. It’s already supported by dozens of blog-related apps.