How exactly will Outbox/Gossip decentralize #nostr if most new users are onboarded to the same handful of default bootstrap relays used by the most popular client apps?
Also, how does search and discovery of content and users work within the Outbox paradigm/approach?
That's the issue, clients shouldn't just all connect to the same handful of relays, they should be able to find the best relay for a particular query and user
The painpoints would be much more pronounced if nostr was 1000x bigger, right now everything works because we are very smol
Already run into issues on amethyst where notes aren't displayed because it's not on a relay I've pre-ser, if it was able to find the relay that note was on, it would be much better
Imho
Default settings in clients are a huge piece of this. Doesn't matter how decentralize-able nostr is if clients push users toward centralization. I for sure see the benefit of outbox enabling clients to find where a note exists, but I don't think it's a silver bullet. I'm really curious about the gaps it leaves in user experience. Not everything works well in a decentralized manner. Search and discovery is one of the first ones that comes to mind.
Ok ok, i'll be serious for a moment. So, with gossip, all you need to do to startup a new gossip client is give it ONE relay. So if there is a relay, that is curated to be a good onboarding relay, you just point a client at that relay.. From there, it will discover all the rest AND give users a good global feed to start with. No brainer.
What about redundancy of user data? One of the selling points of nostr is knowing that if a relay goes down, or if a relay operator decides to shut down their relay for good, that my data exists on other relays. I don't think you could convince me that my notes only should be on one relay.
you could give it two relays. Basically, you should be able to give your friend an onboarding link to a client, and that client will startup with the relays you gave to your friend. That way, they have a less than zero chance of sticking around. Right?
edit: greater than zero
With gossip, your notes don't just exist on one relay.. But the onboarding relay can have a combination of your notes, notes you've replied to, your friends, etc.. Gossip can still publish to all relays if you want, you can even set it up with blastr. It works.
💯
Wondering about this as well. Like Bitcoin and Lightning, most people will take default settings and the easiest path. This isn’t necessarily a bad thing, but it does lead to centralization as the vast majority of users are not going to self host anything unless it’s as easy as pushing a button.
Clients will randomize relay assignments to new users. Primal has already said this will be their plan.
For search, we have different search methods. For those that use a search relay or API, it wouldn't matter I believe because you're not searching just your relays. For Damus that is now using a local DB, that would limit search to just what you've seen, so search wouldn't change here either I wouldn't think? These are both speculation and I could be wrong.
It will be interesting to see how search on nostr progresses and if it can be decentralized. I mean, web search also sucked early on 😂 but web search is centralized to just a couple providers. Can nostr avoid that?
Worth trying. I'm curious how they will determine which relays to rotate through. Also curious if other clients will follow suit or just stick with status quo. Client default settings during onboarding have a huge impact it seems, especially when it comes to the most popular clients.
It doesn't really matter as long as default relay selections meet some threshold of reasonableness. 99% of users could be on the 5 top relays and the network still be meaningfully decentralized. As long as the 1% aggressively banned from the top 5 and the other 10% banned from one or more of the top 5 can still be reached, you're good in my book.
