franzap on Nostr: DVMs are not the perfect design but honestly it's not that hard. Instead of sending a ...
DVMs are not the perfect design but honestly it's not that hard. Instead of sending a REQ you publish a signed request event.
I never tried building a WoT relay so I don't know.
Besides DVMs and the monetization, it makes sense to me for this to be a service rather than precomputed events for the use-cases that clients have (verifying reputation, sorting, recommending follows, etc). Are you going to keep 30382 events for all permutations of pubkeys, for all services, sorted by all different algos, and all reflecting the latest updates to contact lists? A service really is most suited tor this kind of thing, imo.
Also, kind 30382 is too generic and not expressive enough for our needs.
If you'd like, tell me how I would use those to replicate the WoT i currently have on Zapstore (follows who follow). Are you familiar with that? No, it's not recommended follows. It's who of my follows, follow the signer of an app, sorted and limited to top 5.
We can continue the discussion next week when I publish the DVM PRs, as I would love to hear your input and if it could potentially work for your usecase and algos
I never tried building a WoT relay so I don't know.
Besides DVMs and the monetization, it makes sense to me for this to be a service rather than precomputed events for the use-cases that clients have (verifying reputation, sorting, recommending follows, etc). Are you going to keep 30382 events for all permutations of pubkeys, for all services, sorted by all different algos, and all reflecting the latest updates to contact lists? A service really is most suited tor this kind of thing, imo.
Also, kind 30382 is too generic and not expressive enough for our needs.
If you'd like, tell me how I would use those to replicate the WoT i currently have on Zapstore (follows who follow). Are you familiar with that? No, it's not recommended follows. It's who of my follows, follow the signer of an app, sorted and limited to top 5.
We can continue the discussion next week when I publish the DVM PRs, as I would love to hear your input and if it could potentially work for your usecase and algos