---
tags: Meetings
---

# 01/12/2023 Meeting Notes

## Broker Kick Out Proposal V3
explicit about what we defend against (no strategic maliciousness, but majority of users will run standard software)

since they are in same bounty, they already have stake in bounty, more efficient use of capital

brokers should be assigned at time of flagging
SC emits event that these are assigned

smart contract needs to do picking

if validation is intensive and fast -> at flagging
-> if continuous validation is needed then feasible to choose at join time

eth blocks might come with randao to use (CHECK)
if you grab block form json rpc -> Randao can be used by SC?

everyone whos adress hashes to this range is chosen?

enforce broker contract ->
factory of brokers -> can arrange to know set of all brokers (or the number)
brokers need to know whether they are assigned, and contract needs to verify
can always join bounty if you want to cheat
global list
selection of brokers: pick voters that are live (look at all broker contracts, some could be abandoned)
if broker is in bounty -> should be active

voting is essence of problem
how to conclude vote (no timer)
-> secret voting protocols
-> others should not know what the current vote tally is
-> needs another reveal tx (if done with pre-commit)
secret sharing? pieces of puzzle?
necessity -> closed voting

closing and resolving after required votes are in would be nice (pick 20 require 7 -> could close if 4 yes)

removing freeriders is already valid incentive for flagging, so the payment for a valid flag could be 0
stake to flag needs to be balanced correctly, payments too
cost to do validation potentially low, no need for high payments / stake

process has to be automatic
in miner node -> only attackers will modify, vast majority will run default
-> could be ok to have several txs, cost is mainly implementation

traffic relayed to validation aspect all the time with people watching each other
chosen brokers will have a way of forming opinion yes or no

remove unnecessary parts, iteration
sponsor rights adds more problems than it solves (can become a sponsor with min funds and just veto stuff)

Collaboration
brings us to end of track with BlockScience
what model do we have for working with us
on-demand would be great to feed in workstreams instead of regular
once implementation done revisit digital twin / CadCad implementation

late addition:
timing wise: might want to do -> flag raised -> stake put up -> brokers assigned -> BREAK (1h) -> Voting can start
this makes sure that people don't start randomly voting to have their vote counted for sure (would need to deal with early voting visible in mempool though?)
also need to think about how long a proper validation takes (and for how much longer we want to do validation to decrease the risk of punishing someone who has power outage -> already decreased by brokers probably not putting up stake for peer review if someone is unresponsive for an hour, they would probably only do so if they're really sus of that other broker)

## Broker DD V3