The only thing they seem to miss from the author's description is: "It has to be decentralized. The database would be hosted on many computers and each new entry would be processed using some sort of consensus algorithm."
But there's no motivation for why this is a expected property of the network
RSS is a good example. In the simplest form my proposed social network structure could be seen as RSS feed of all the social web that is stored in a decentralized database. The main feature is that the posts can be 'liked' or 'upvoted' by users (domains).
The sudden idea is to separate metadata and content on different services and merge them only for viewing and interaction. Then users can selfhost content and request its distributed social metadata (upvotes, emotes, other people's comments etc) dynamically without easily losing control over the content itself.
Maybe, but the article needs to talk about that. That is my point. From my own view, I'm using both of those things and it does work, so why would I entertain a slough of new ideas to be started from scratch without seeing a compare and contrast discussion?
Oh god, someone posts this sort of thing every couple months. Also tends to have high correlation with crypto-folk.
Decentralized social networks are a pipe dream. I'm sorry but no one wants to join you and your 50 shady hacker friends in a forum.
People want to go where their friends and family are. People's friends and families aren't going to figure out whatever stupid, inconsistent, under-designed UX people inevitably build for these sorts of things. The bar is extremely high.
Ideology seems to come before any practical notions of how people use products.
This is true and remains true, but if it ever changes and the UX becomes un-under-designed, I don't see any reason we can't proceed. Keep an eye on Matrix as they're undergoing a UX revamp and the people working on it are super thoughtful.
We have a decentralized public social network: blogs. Webmention and microformats2 standards build on top of blogs to add federated mentions, replies, likes, etc.
20 comments
[ 3.3 ms ] story [ 51.2 ms ] threadThe only thing they seem to miss from the author's description is: "It has to be decentralized. The database would be hosted on many computers and each new entry would be processed using some sort of consensus algorithm."
But there's no motivation for why this is a expected property of the network
They "work" in the sense that they have a checklist of features that I'm sure function as intended.
But the product/network as a whole is far from "working".
In what way?
Decentralized social networks are a pipe dream. I'm sorry but no one wants to join you and your 50 shady hacker friends in a forum.
People want to go where their friends and family are. People's friends and families aren't going to figure out whatever stupid, inconsistent, under-designed UX people inevitably build for these sorts of things. The bar is extremely high.
Ideology seems to come before any practical notions of how people use products.