Topic: Kudos: A Peer-to-Peer Discussion System Based on Social Voting (Read 1306 times)
You should check out http://coinawesome.com/ and come back to me
Abstract
We describe a peer-to-peer discussion system that uses a blockchain for tracking users' points which are mapped to specific content, and a distributed hash table for storing/serving content. Points are minted through the mining process which achieves consensus on the chain to accept. Peers only accept transactions (aka votes) of a universal constant amount. Fees for the miners are also constant to prevent self-promotion. Users earn points either by mining or by posting relevant content that others are willing to upvote. Highest blocks in the chain contain references to the latest most relevant content, which is republished more frequently by peers to enable quicker response times. The system has the property of allowing peers to discuss openly and freely, without the fear of being blocked or banned. It also allows relevant content to stand out without the need of a central tracker or index. Peers gain reputation through the points they earn but also through the irrevocable proof of the things they said available in the blockchain.
DRAFT paper: http://lucaa.org/docs/kudos.pdf
---------
I've had this idea for a very long time now and finally decided that the "paper" is somewhat in a good enough shape to post on here and hopefully get some feedback. Please understand that it's still very much in DRAFT - lots of things are missing and I decided to publish it in this state because I'd like to know whether there's something fundamentally broken that I'm missing from my design before I move on in the implementation.
In essence it's an idea to rank *digital things* using a blockchain. Scarce tokens in the blockchain are used as points that people give to one-another, and each transaction contains a reference to these *digital things*. So one can calculate the relevance of specific things by simply looking up the transactions in the blockchain. It's interesting because we don't really have many ways to rank things in an entirely decentralized way.
Use cases:
* Obviously a decentralized Reddit is my main use case. Comments and posts are stored in the DHT. The blockchain would only contain the "key" (of the hash-table key-value pair). I join the network, and download latest blocks. Since each "uvpote" is essentially a transaction, I can calculate the relevance of specific DHT keys. Then once I have all the relevant keys I simply get() them from the DHT network.
* Torrent search-engines are obviously one major point-of-failure for the torrent ecosystem. Imagine having to simply look up latest blocks in a blockchain to figure out which torrents are trending at the moment from a specific sub-group.
* Proof of what anybody said. Imagine having a lengthy discussion with someone and wanting to prove this discussion took place. Since all the records are kept in the blockchain you can not only prove someone's reputation, but also that they said something in a specific moment in time.
Well that's it for now. Let me know what you think!
We describe a peer-to-peer discussion system that uses a blockchain for tracking users' points which are mapped to specific content, and a distributed hash table for storing/serving content. Points are minted through the mining process which achieves consensus on the chain to accept. Peers only accept transactions (aka votes) of a universal constant amount. Fees for the miners are also constant to prevent self-promotion. Users earn points either by mining or by posting relevant content that others are willing to upvote. Highest blocks in the chain contain references to the latest most relevant content, which is republished more frequently by peers to enable quicker response times. The system has the property of allowing peers to discuss openly and freely, without the fear of being blocked or banned. It also allows relevant content to stand out without the need of a central tracker or index. Peers gain reputation through the points they earn but also through the irrevocable proof of the things they said available in the blockchain.
DRAFT paper: http://lucaa.org/docs/kudos.pdf
---------
I've had this idea for a very long time now and finally decided that the "paper" is somewhat in a good enough shape to post on here and hopefully get some feedback. Please understand that it's still very much in DRAFT - lots of things are missing and I decided to publish it in this state because I'd like to know whether there's something fundamentally broken that I'm missing from my design before I move on in the implementation.
In essence it's an idea to rank *digital things* using a blockchain. Scarce tokens in the blockchain are used as points that people give to one-another, and each transaction contains a reference to these *digital things*. So one can calculate the relevance of specific things by simply looking up the transactions in the blockchain. It's interesting because we don't really have many ways to rank things in an entirely decentralized way.
Use cases:
* Obviously a decentralized Reddit is my main use case. Comments and posts are stored in the DHT. The blockchain would only contain the "key" (of the hash-table key-value pair). I join the network, and download latest blocks. Since each "uvpote" is essentially a transaction, I can calculate the relevance of specific DHT keys. Then once I have all the relevant keys I simply get() them from the DHT network.
* Torrent search-engines are obviously one major point-of-failure for the torrent ecosystem. Imagine having to simply look up latest blocks in a blockchain to figure out which torrents are trending at the moment from a specific sub-group.
* Proof of what anybody said. Imagine having a lengthy discussion with someone and wanting to prove this discussion took place. Since all the records are kept in the blockchain you can not only prove someone's reputation, but also that they said something in a specific moment in time.
Well that's it for now. Let me know what you think!
Jump to: