Oddbean

So Web5 has completely ditched ION and don't even explain to anyone why?, or even make a comment about ditching it?, they will just pretend ION never existed after so many years of telling all the critics that ION was this amazing idea and that anyone who dared to criticize it was dumb? https://twitter.com/csuwildcat/status/1761168521663316228

The most stupid thing about the entire concept of a "DID" is that it was supposed be this one interoperable thing, right? You use "DIDs" in your project and now people using all sorts of DIDs can interoperate. But what actually happens in practice (obviously) is that projects have to pick one specific DID and support just that. So Web5 only works with "did-dht" and Bluesky with "did-plc" (aka their server) and one is not compatible with the other. But somehow Nostr is less compatible and morally wrong because Nostr keys don't have a "did:" prefix in them. nostr:nevent1qyd8wumn8ghj7urewfsk66ty9enxjct5dfskvtnrdakj7qgmwaehxw309aex2mrp0yh8wetnw3jhymnzw33jucm0d5hsz9mhwden5te0wfjkccte9ehx7um5wghxyctwvshsz9thwden5te0wfjkccte9ejxzmt4wvhxjme0qy88wumn8ghj7mn0wvhxcmmv9uq36amnwvaz7tmwdaehgu3wvf5hgcm0d9hx2u3wwdhkx6tpdshszxrhwden5te0wfjkccte9e3h2unjv4h8gtnx095j7qpqzxjq7cesh9dehesk60k99xlgkqf4st98l8ax2kz6wzd4zt08nkeqz3ts7a

Yep, DID Resolvers have always been a huge issue. Funny thing: the universal DID resolver https://dev.uniresolver.io/ instantiates a docker machine for each DID method because it's impossible to develop a single library that resolves them all.

Why would a federated RSS feed be 'the beginning of the end' for decentralized app datastore tech (think: decentralized version of Firebase)? Those two things are wildly different, so I'm not understanding your comment - it's like saying 'the automobile was the beginning of the end for the airplane'.

I was more asking why you think Nostr possibly succeeding in the future means that we don't need other technologies. These two technologies, analogically are basically RSS : Firebase :: Nostr :: DWeb Nodes, so it's not like you can build every type of app on RSS, you eventually need datastores. Why do you view this as some battle where there can only be one winner? Did cars make planes unnecessary for you?

So for the signing in process with username and password you need to pick-up a specific DID from the very start. Otherwise if only Nostr protocol is used, users can sign in using nsec only and there won’t be a way to integrate any other signing in method further ahead because of the incompatibility with DIDs. What’s used in Alby?!

They gave up on ION because it has always been moronic idea, just like did:plc They eventually gave up and built a DID on Pkarr.org instead. nostr:nevent1qqsprfq0vvctjkumuctd8mzjn05tqy6c9jnln7n9tpd8px639hnemvspp4mhxue69uhkummn9ekx7mqzyqalp33lewf5vdq847t6te0wvnags0gs0mu72kz8938tn24wlfze6qcyqqqqqqgzzr5mt

If you remember, I told you I will build something that uses Mainline to point to anything like DNS, that became Pkarr.org that did:dht is based on. If you are open to adopt it despite not being secp256k1, let me know. Also check what Iroh wrote about it iroh.computer/blog/iroh-global-node-didcovery

We’re not going to completely change our keys and identity system to get a system to look up name/key to relay lookup system. Doing that would mean forking Nostr. You’re welcome to try and fork the tech and community, but that’s not an ocean I want to boil.

That is perfectly reasonable, but my point stands that the only alternative is to use a DHT that supports secp, and then all your choices are going to be so small that censorship resistance isn't as credible as Mainline. There are creative ways to use Pkarr in backwards compatible ways, similar to how Nip05 is optional, but it is not a thing I will personally propose because it will be perceived as hostile NIP probably.