@manan

on adding more validators - our criteria for adding validators right now is simple: run snapchain, not be neynar and be in a geographical region different from the current validator set. why? prioritizing stability, decentralize 1 variable at time. top 2 priorities in order right now - geo and no majority. ---------- we want more validators and to move farcaster to be decentralized. we were advocating for it when we were running neynar's own validator node before the acquisition, during the transition, and once the deal had fully closed at the beginning of april. it's been a consistent position. what we can't do right now is onboard a validator running a modified version of snapchain. snapchain is a live consensus system. a validator running a fork, even a well-maintained one, creates risk we don't currently have systems set up to verify from the outside and don't have the dev capacity to take on right now. when something goes down on any part of the validator set, yes it is largely on the team running the validator node to address, but as the team stewarding the entire protocol, at the end of the day we're the ones accountable for the network. on the question of compatibility, yes, we've merged contributions that came through PRs from outside the core team, including from cassie. that's how open source is supposed to work and we're grateful for it. but merged contributions aren't the same as a guaranteed-compatiblity. we've also significantly edited and rewritten portions of what was submitted. the codebase running on mainnet today is not a 1:1 copy of any external PR, which is part of the work that we were willing (and excited) to take on as stewards of the protocol. honestly, this is part of why we encouraged alternative paths in the first place. we could see that our timelines and priorities weren't going to line up, and if someone wanted to move faster or in a different direction, experimenting in a different direction was the right path. that's how open protocols are supposed to work. we meant that genuinely. we do want farcaster to be fully decentralized eventually. we just disagree on the order of operations to get there. stability and growth of the network has to come first - a decentralized protocol that nobody uses doesn't serve anyone. to reiterate, our criteria for adding validators right now is simple: run snapchain, not be neynar and be in a geographical region different from the current validator set.

11 replies

9 recasts

47 reactions