mleku on Nostr: there has been some chatter recently mostly involving #gitcitadel people talking ...

there has been some chatter recently mostly involving #gitcitadel people talking about the fact that nostr relays are essentially distributed database servers, of a new kind that pushes more of the subscription and authentication part to the front

i'm of the opinion that as shitcoins slowly die, because they are over-replicated distributed databases without a built in subscription protocol, that a nostr-like distributed data protocol is growing out of the nostr design

bitcoin will become the authoritative anchor for these decentralised data stores, and small, low bandwidth, ad-hoc federated blockchains will become directory servers, that use bitcoin as their security, anchoring their state in short messages stored in op_return, or maybe witness data, or both, and these will become directories that store information about where categories and users (as in decentralised web services) are storing their data on a growing network of nostr-like relays

nostr protocol will have to be refined for this use case, as it was originally conceived primarily as a decentralised replacement of the centralised, distributed data systems used by current generation social networks. they will need a simpler interface (ie, HTTP based) and where currently there is one spec for data, there will evolve a simpler, coordinating data storage interface that treats database records as events and relays will have user-programmable replication schemes where the relay acts as a client to other relays, and using auth-proxying and data storage proxying protocols, will become the place where all kinds of data are stored, protected by auth-required access, and enabled by clients being able to use relays as proxies.

probably the main thing that i see coming is that nostr relays will become partially also proxies for clients to fan out their requests to multiple data stores, providing a replication state programmable by the smart contracts running on these low bandwidth, data-only pBFT blockchains, and client applications, that read and write to the relays, with the blockchains becoming the directories.

i know it probably will be difficult for most to understand what i'm talking about, but the reason for this to become the dominant paradigm of internet services is that it separates application logic and data storage logic, and enables a continuum of scaling that has low friction for applications that become used a lot, to scale up their back end without having to change any of their application logic.