Bluesky the company is spinning off the the name system, DID:PLC in to an independent swiss association.
https://web.plc.directory/
https://docs.bsky.app/blog/plc-directory-org
This is a good step, the name system should not be owned by the company.
Login to reply
Replies (16)
Those DIDs should be distributed. They know that right? π§
Well originally PLC stood for placeholder but when it stuck they renamed it "Public Ledger of Credentials". At least they're not moving it to an independent organization. Not perfect, but a step in the right direction.
Step by stepβ¦
You know that there is did:nostr spec (draft) too?
https://github.com/nostrcg/did-nostr
it's not supported by any apps and is kind of parallel to nip-05.... Tech adoption is only a little bit the tech itself... it's a lot of the social work of getting something working.
π―
Real adoption comes from solving real problems. Not easy.
You have to make some trade-offs to compete with centralised services and their traditional UX elements. Stuff just works on atproto, it's all smooth and consistent and familiar to normies.
This feels like them just trying to balance the trade-offs.
Nostr as a protocol cares relatively little when it comes to these traditional UX elements and so it's much freer when it comes to identity.
I agree, they made choices that worked and they were able to onboard a large user community.
Still a placeholder, because without DIDs, the whole system isn't more decentralized than ActivityPub, which is already a widely used standard now.
In fact, if all records are centralized in a single organization, it's much less decentralized than any normal federated protocol, because it creates a single point of attack. How is that a step in the right direction? Genuinely interested in why you think that's good.
I think it's better than centralized and controlled by the company. little steps...
But it's still completely centralized, no? So the right direction (in my book) would be towards decentralization, not towards ossifying even more centralization.
But with Nostr we can make the same trade-offs right? And build a Bluesky app in a very centralized way and provide that smooth UX. Somewhat similar what Primal is doing.
Be careful, NIP-05 is just a way to define some DNS alias to a npub. But the identity itself relies in the pubkey/noun, never in the DNS alias.
I can change the DNS alias, and it would still be my identity if I keep my npub.
But if I change the npub and keep the DNS alias, that should break things (WoT).
π«
In theory maybe, but in reality Primal's UX is an order of magnitude more clunky than even an indie atproto client thrown together by a single dev in a few days. Many things are touch-and-go on Primal, and common traditional UX features are missing outright.
This is because Primal still has a long and messy list of tasks it has to get done in the background and it has to do all of this itself, it can't rely on shared protocol infrastructure that makes things easier and smoother for everyone.
For Primal to get to the level of smoothness and consistency of even the most thrown-together atproto client of today would take years, and the result would be absolutely massive cloud costs versus what the atproto client is paying today.
Still centralized. And the protocol's userbase is still like 99% Bluesky.
Depends how you look at it. It's a directory, at the end of the day. The ultimate authority is the DID itself not the directory. The DID itself is derived from a hash of the initial operation signed with the private key.
Anyone can mirror that directory. Because each op is cryptographically signed, anyone who mirrors it can validate the entire log independently. Or you can have multiple, independent groiups operating replicas of the directory.
But because of their UX you need some source of truth, hence it's not completely centralised but still not able to be completely decentralised either.