Solid results, this is really good! I notice however that it relies on relay.nostr.band to actually find the event ids the DVM returns 😂 The DVM return format should include relay hints nostr:nprofile1qy08wumn8ghj7mn0wd68yttsw43zuam9d3kx7unyv4ezumn9wshszrnhwden5te0dehhxtnvdakz7qg4waehxw309aex2mrp0yhxgctdw4eju6t09uqzpxdm2kgujytxqruy2yrax8umt83003lqng0lsqhgfuw58kj40jnytw4yre
Yes, but this is a proof of concept. There is currently my DVM and that relies fully on nostr.band (and a second version that gives up to 150 results from a year ago, also from nostr.band). The intention is that anyone can run a DVM and improve results further :). The page collects all events it gets.
In terms of relay hints, please feel free: :)
https://github.com/nostr-protocol/data-vending-machines/blob/master/kinds/5302.md
Thinking about this a bit more today and have a few questions.
In an ideal future, who do we expect to run the search DVMs? If it’s aggregators, I’d assume the relay hint would always be their own relay since they may not store the original source relay of the event and do not know if it’s even still available there. If it’s individual relays operating the DVMs, why not just use NIP-50 where the full events can be returned in one REQ?
I think it will depend on how many relays we have. If we go for thousands or millions of relays, then I do believe DVM service will be needed.
It's a wonderful POC. I've been planning to write a similar thing, but for profile search.
Somehow I thought I invented 5402, which is similar, but I don't see the PR in the repo 🤷♂️. Maybe I never pushed my branch. DVMs for second-layer nostr functionality are a great area of research.
Havent seen a 5402 so probably not pushed.
Absolutely. We're only beginning to explore what's possible.
