After half a decade of #nostr we still cant create an account on one app, log in to another app, and follow someone. Try it, it's buggy at best, and almost impossible at worst.
We really need to fix this because multiple apps working together it's Nostr's super power.
Wdym? Yea you could.
To use the outbox model or not to use the outbox model? The answer to that question should be uniform, in either direction. That would improve interoperability a lot.
Outbox is a bit of a "No True Scotsman" fallacy. It sets up an unfalsifiable position: whenever there's a problem, people say, 'Ah, you just didn't use Outbox correctly.' That risks creating cargo-cult thinking, where everyone copies patterns without really understanding why.
The thing you need is consistent profiles. And I am unsure outbox really helps here.
I've used both models, and honestly, the outbox model just makes everything overly complex, and gums everything up if you're switching between the two models using the same nsec. In my opinion, it's a bad idea to tell pre-Nostriches they can take their nsec and use it with any Nostr client; then, when they discover that's not altogether true, we've lost them.
Elephant in the room is that #nostr is not a database system, it's a relay system. But no one wants to admit it.
Seems not that expensive to crawl, index and firehose-ify pretty much all of Nostr in its current state. As far as I know Primal's doing it, Ditto's doing it, Nostr.band's doing it, and maybe Nostr.wine's doing it too. Who else?
That trends spreads it'll become a globally-indexed relay system with multiple global indexers.
I'm also not sure that's a bad thing. Or a good thing. Have to let it seep in.
It's a good thing!
I think it could be. "Index Nostr" takes a lot of pressure off "Relay Nostr" and vice versa. But then the focus should be on smooth switching between the two Nostrs, as well as graceful fallback from Index Nostr to Relay Nostr when needed.
In that world I suppose Outbox (a feature of Relay Nostr) becomes at worse a waste of time, and at best a way for Relay Nostr to provide some of the features of Index Nostr in the event that Index Nostr goes down. But with multiple indexers and the relative simplicity and low cost of crawling and indexing the Nostr universe, is that really necessary?
Relay Nostr without Outbox still provides decent fallback in the event that Index Nostr somehow goes down all at once, or for users who want a pure relay experience, or for hardcore censorship resistance use cases. So Relay Nostr without Outbox still takes pressure off of Index Nostr in the sense that it makes Index Nostr less of a scary centralising force.
And it's unique. ATProto doesn't have this relay mid-layer—it's either pull from the firehose or pull straight from the PDS. That relay mid-layer (or fallback layer, depending how you see it) seems like a real architectural advantage for Nostr.
I suppose if everyone accepted this Big Nostr / Little Nostr future then we could go full speed ahead on architecting graceful interaction between the two. Many things now taking a lot of time and effort would instantly become distractions that could be let go of and forgotten forever (mainly painful efforts to squeeze a few drops of index juice out of the dried up relay orange).
I wonder where the current Nostr zeitgeist lies on all that? Anyway with Primal, Ditto, Nostr.band, Nostr,wine, et al. doing what they're doing seems like a hefty part of Nostr is prepping for that dynamic regardless.
Definitely a good thing!
Outbox model is shit because it's user based rather than content based. I don't care about following some user I already knew about talking about new topics, I'm interested in finding new users posting content I was already interested in. I'm pretty sure there is no way to do this without new relay transparency measures.
