yeah, nostr needs a credible upgrade. one where a proper HD keychain enables revocation and doesn't put the master revocation key pubkey all over the place, just one signed revocation that once you publish it it advances your identity using an xpub and everyone can find it automatically because it's an xpub, but it's down stream of the HD path. that would fix that entire problem in one fell swoop. put bip-39 word key on the seed and it's a memorisable keylock and an extension or native app bunker style signer becomes viable. for native apps, that bunker can have localhost access, the web app path unfortunately that's complete bullshit and crap that you can't work around (the evryfing must be SSSLLL bullshit).
and the encoding for messages? json is total bullshit. it should use email style start of line sentinels (also, Go, Pascal Modula and Oberon style code block sentinels) and... well, these are my two gripes with nostr. aside from that, everything else about the protocol is awesome.
oh yeah, one more gripe. kind numbers. fuck that shit. it should be fully qualified path based identifiers. textnote. someprotocolname/someprotocolmessage. and so on.
yes, ok, those 3 things fixed and it's alright. everything else is just a matter of client and relay stuff. i mean, auth is required. you can have subscriptions that let you open protocol message tunnels already, that's what bunkers already are, they could be a path to location concealment with the use of nwc and keysend. running relays is trivial, there could be dozens if not hundreds of nostr relay hosting services that you can just throw a few dollars a month and badabing badaboom you have your own trustworthy or more trustworthy rendezvous points that get around the CGNAT cage of the modern "users can't run servers" dictum that is maintained mainly by fucking mobile phone company lobbies.
