Feels like the A.1 issue of Mastadon as a platform. If person A on instance Q wants to follow person B on instance R, there’s no straight line easy path to do that. Compared to Twitter or BlueSky or Threads, where its all one ecosystem and you just say “I’d like to follow @LieutenantDickweasel” and now you’ve got their posts in your stream, Mastadon is byzantine and not worth the effort to explore.

You do know that the Fediverse is more than just Mastodon, Truth Social and the Threadiverse?

Search that covers 100% of the Fediverse is technologically impossible. Any Fediverse-wide search would need to know all of the Fediverse. All of it.

Like, let’s suppose R is B’s personal instance. Let’s suppose B spins up the instance for the first time. Any all-encompassing Fediverse search would have to know about it immediately. The very millisecond Apache or nginx or whatever comes to life, that search would have to know it’s there to be able to always cover exactly 100% of the Fediverse.

How’s that supposed to work?

If it’s one centralised search engine, it would have to be hard-coded into the source code of every last Fediverse project out there so all new instances can automatically announce their existence to the search engine.

And that’s not four projects or so. It’s over 100. Not only Mastodon, Lemmy, Mbin and PieFed. It’s also Ecko and Hometown and Glitch and many other Mastodon forks. And Pleroma and Akkoma and other Pleroma forks. And Misskey and Firefish and Iceshrimp and Iceshrimp.NET and Sharkey and CherryPick and Catodon and Meisskey and Tanukey and Neko and dozens upon dozens of other Misskey forks. And Mitra. And Socialhome. And GoToSocial. And micro.blog which, by the way, is closed-source. And Friendica and Hubzilla and the streams repository and Forte. And Pixelfed. And Funkwhale. And Bandwagon. And Castopod. And PeerTube. And Owncast. And Mobilizon. And Gancio. And BookWyrm. And Flohmarkt. And so forth.

It’d be even worse if it was supposed to be built into the Fediverse projects themself. Like, you could search the whole Fediverse from Lemmy’s Web interface or any one Mastodon app.

That’d require each new instance to announce its instance to each running instance.

That’d require each new instance to know all running instances immediately.

That’d only be possible by building a list of 20,000++ Fediverse instances into every last Fediverse server software repository so that it’s installed along with new instances.

And that list would always have to be up-to-date.

So when B spins up R, the following would have to happen:

• R git pulls the most recent version of the main branch of Mastodon’s source code to have a most up-to-date list of active instances possible.
• R starts up.
• R announces its existence to the 20,000++ Fediverse instances on the list.
• R goes through a list of all Fediverse server application code repositories which it has pulled from the Mastodon code repository as well.
• R announces its existence to every last one of these repositories by creating a new branch, editing the list of active Fediverse instances, submitting the edit as a pull request and merging its own new branch into the main/stable/release/… branches of all these code repositories.

Any Fediverse server out there would be able to hack into any Fediverse server code repository and manipulate the production code. Otherwise, this whole thing couldn’t work.

Fediverse server code repositories would be flooded with automated pull requests plus mergers. Oh, and if Mastodon can add a new instance to a list in the Mastodon production source code, anything could remote-manipulate anything in the Mastodon production source code.