At its core, Before the Ban implements a decentralized web of trust based on assertions and witnesses related to the connection between social media users, individual persons (or personae), and content URIs. The web of trust is overlaid on existing social media networks, and “backed up” to decentralized storage through IPFS. All assertions and witness are, in effect, sets of vertices and edges in a graph database. Each of these “claims” has a globally unique ID related to the hash of the data, the user’s public key, and their signature validating possession of the corresponding private key.
Before the Ban uses a trust ranking algorithm that takes into account the number of witnesses for an assertion and the “quality” of those witnesses. This algorithm resembles the original PageRank system used by Google.4
However, a key strength of Before the Ban is that the trust ranking algorithm can be swapped out or forked at any time, by anyone. This should reduce the incentive to game the trust rankings, as there will be an unknown number of ranking systems in use, each one emphasizing different factors and evolving over time to overcome shortcomings.