Conversation

Notices

1. Embed this notice
   Sick Sun (sun@shitposter.world)'s status on Monday, 18-Nov-2024 22:25:45 JST Sick Sun
   I should like Nostr a lot because I like the idea that as much is pushed into the client as possible so it can be decentralized as possible but in practice I feel like they deliberately avoided learning anything from what came before it and every design decision makes me cringe
   • soberano likes this.
   • Embed this notice
     Niggy Hardare (j@cawfee.club)'s status on Tuesday, 19-Nov-2024 00:32:58 JST Niggy Hardare

     in reply to
     @sun You say that a lot but never give specifics.
     
     I soured on NOSTR's design bc in practice the entire thing is centralized on a couple of relays that are virtually duplicates of one another. You can use alternatives but at the cost of not seeing most of the network.
     
     It's not very decentralized and it's radically inefficient. You get portability but you lose so much more. It's inefficient to the extreme.
     
     Hopefully something better comes along.
   • Embed this notice
     Sick Sun (sun@shitposter.world)'s status on Tuesday, 19-Nov-2024 00:32:58 JST Sick Sun

     in reply to

     • Niggy Hardare
     @J they create their own endpoints using their own formats for linking domains to names instead of using .well-known and webfinger or something. event construction is a mishmash of object properties and flattened properties in arrays. this mess is done to avoid having a canonicalization scheme. the specific way they did it relies on position inside arrays, it's just really fucking bad. the event ids, if you want interop, rely on a centralized process of recognizing event ids instead of using like namespaced strings or hash of a deterministic layout spec. when they define new events a number of them hardcode lists of services or functionality so now for eternity you have to just create new proposals to add to that list long after a service it mentions could possibly die. events for things like reporting content have a hardcoded list of reasons instead of configurable and the list was made by a stupid person
   • Embed this notice
     Sick Sun (sun@shitposter.world)'s status on Tuesday, 19-Nov-2024 00:39:47 JST Sick Sun

     in reply to

     • Niggy Hardare
     @J yeah my feeling is that they applied a series of really naive approaches to things like uncensorability and then punt on making them work well and furthermore aren't even ready for serious censorship attempts. like they made a spec that pushes the hard problems into the clients and brag about it like theyre geniuses, while the actual clients don't work so great.
     Linux Walt (@lnxw37j1) {3EB165E0-5BB1-45D2-9E7D-93B31821F864} likes this.