the web was not made to be federated imo because the web is not a messaging platform, it's a publishing platform. in messaging, you federate between systems; in publishing, you *syndicate* to other sites. there is a difference, and the difference matters.

@trwnh but at the same time the web is kind of just the default application platform now. so if you wanted to build a message passing system you'd be expected to do it using the components of the web, http as tcp/ip, even if it means wrapping streams in POST requests

@alice this is true but i was thinking more about the architecture of something like the fediverse being somewhat of a misnomer or otherwise just ergonomically awkward — the protocol works on a messaging basis but we are using it in a publishing context. among other muddiness…

@alice maybe it would be a bit clearer to say that instead of sending a message to followers, we are publishing our posts to every single site that hosts at least one follower. it’s not like fedi software is generally exposing inboxes directly to users as if it were email. instead, every actor is just being puppeted by the “instance” and is behaving programmatically in response to user actions, rather than letting users compose their own raw messages

in messaging the final target is the recipient, in publishing the final target is a resource.

here on fedi we're less "sending messages to followers" and more "publishing a syndicated copy of our posts on every site that hosts at least one follower". as a general rule, fedi doesn't let users compose raw messages, nor access the raw inbox. all interactions are mediated through actors being puppeted by the "instance", targeting syndicated resources that are at best manipulated through messages.

@trwnh yea but that's more of an implementation detail that stems from constraints of the tools available (the web), than really the purpose of this nebulous mess of protocols we call the fediverse? if most use it for mostly-private messaging, then maybe it's not about publishing at all, or we would indeed be writing blogs.. but not exactly. i think there is absolutely a duality at play where there are two expectations: 'are you using this as chat app or a blog app?' basically in our old words, and implementation that are very unsure which they should cater to or try to cater to. or makes it a 'privacy level' switch and then tries to fill in the gaps as they come

@alice there's definitely a duality and imo it ends up just leading to Problems to not address it and pretend that it's not there. but some of it goes deeper than the ux level, some of it is about the fundamental paradigm, the "shape of reality" as it were. you could do some form of "publishing" by just mass-messaging everyone in your contacts, like a newsletter or 2006-era twitter. it would work but it would be awkward bc you can't easily "unpublish"; a message sent is out of your control

@trwnh could you expand on the difference? how would this play out?

@okokokok sure! so if we leaned more into publishing, then we might have features like being able to track which sites any given post is published/syndicated to. my canonical profile on mastodon.social would be linked to a syndicated profile on mograph.social. they would be different profiles, but each profile would declare that the person it represents is the sameAs the other profile. if i had some way to prove identity, i could “claim ownership” over my profile on mograph.

@okokokok if we leaned more into messaging, then we might allow users to browse their inboxes just like they can in email. users could send messages to other users in the form of activities or notifications of whatever they are doing or did. users could even send messages to applications or services, and depending on the shape of the message, it might trigger certain actions. “i listened to this song” or “i traveled to this place” or “i like this thing” would all be examples of messages.