Also if you see a specification being broadcasted on a relay you can expect there's a chance that it is being used on those relays. Could reinforce territorial dynamics though.

But beyond that, a registry can stand as endorsement of the content, and endorsements of the specifications it links to. Other users endorsing that article can gauge the level of concensus for specifications. Rather than a specified length, it will decay for less known/used nips,

Then there is Vitor's demo of running a webserver from nostr events. How can you determine the trustworthy events on a relay where anyone can send events? You can read a registry from a trusted user (at this point, i mean actual trust, not computational WOT stuff) and that can act as a dns server to other nostr events. If we were to computationally do it, we'd need a strong metric that cant be gamable. Otherwise I'd see how that'd be a big safety hazard.

nostr:nevent1qqsdzdgspd8tdw8w2c6vnkeq206gwpsk5k0j99757t205hg7jdykghgpr3mhxue69uhkummnw3ezucnfw33k76twv4ezuum0vd5kzmqzyrwye5yxe47wtvvr9t05lhgjzy5f3qxjcl3ft09su6zvqxkwua7qvqcyqqqqqqgcghnqt

Reply to this note

Please Login to reply.

Discussion

I guess I add a layer of "human trust" because I'm not going to merge anything I haven't examined.

That wouldn't block someone from adding his to an additional page, like an addendum, but it probably wouldn't be auto-extracted and used in a relay or something. He'd have to market it, separately.

Bring back old school human trust to technology and the internet! nostr:npub1f6ugxyxkknket3kkdgu4k0fu74vmshawermkj8d06sz6jts9t4kslazcka might be an interesting angle from a security/privacy aspects about this and how we can work through them.

Specifically:

We can formalize the specs we recognize, we can signal users to be trusted to not run malicious source code. Using human trust, you can at least say that code from their events is less likely to be malicious. Computational trust, we'd need a hardened method for that.

I mean, the holy grail of event registries is a "used in client" field. Which would also make obsolete events obvious, as no client would be using them.

But the biggest change is just the recognition that the event registry is a dynamic document worthy of its own independent management, versioning and co-editing, rather than being buried in a proprietary git repo.

Not the least because it's the most-powerful free marketing.

I love the scaling dynamics. Keeping versions is optional for you, you can keep however many versions you want for your data needs. Though, if your ideas/specs/product are popular, they'd be cloned to keep a backup of. The trust dynamic would work that if one of your endorsements ever become malicious, you are incentivized to remove them asap. Or move to another list - known malicious users.

Could maybe be a separate list, at the bottom of the page.

Otherwise, people wouldn't find them by searching the page and they might not know that there's a second page.

What about using a traffic-light system for the non-malicious list. Something like:

🟢 active events

🟡 obsolete events

🔴 deprecated events

Ux needs work, but i think we can fix this with "nostr register" or just wiki entries with endorsments of clients or npubs

nostr:nevent1qqsdc8mjyxjefrlxjatsvt395eqx942j0fam9zdfw8wdgtzatqvfwqgpz9mhxue69uhkummnw3ezuamfdejj7q3qm3xdppkd0njmrqe2ma8a6ys39zvgp5k8u22mev8xsnqp4nh80srqxpqqqqqqzt2fuqj