Did you see the Discovery Servers?
> A self-hosted, headless server binary (server/) that adds opt-in aggregation, search, and trending to the P2P network. Users register with a server by signing a ...
I actually worked on this stuff and thought about it deeply, but I am not very optimistic on it at scale.
Pushing notifications to an aggregator whenever you have something new, the ActivityPub way makes sense. And Crawlers make sense.
But pushing a "hey please subscribe to me" the way Bluesky does is weird although I actually designed Homeservers to support that. It only makes sense if you assume orders of magnitudes fewer Homeservers than users (little self hosting) because it saves on bandwidth, BUT if you have a PKI, and you don't support random DIDs some of which can't be enumerated, then you could just enumerate all users and thus all their Homeservers and thus no need for the Homeserver to notify the aggregator of its existence.
I don't know, all of it seems marginal to crawling
Pushing notifications to an aggregator whenever you have something new, the ActivityPub way makes sense. And Crawlers make sense.
But pushing a "hey please subscribe to me" the way Bluesky does is weird although I actually designed Homeservers to support that. It only makes sense if you assume orders of magnitudes fewer Homeservers than users (little self hosting) because it saves on bandwidth, BUT if you have a PKI, and you don't support random DIDs some of which can't be enumerated, then you could just enumerate all users and thus all their Homeservers and thus no need for the Homeserver to notify the aggregator of its existence.
I don't know, all of it seems marginal to crawling
1