Hi all, given the market conditions over the last week I wanted to open the discussion about speeding up on-chain governance decisions – at least temporarily. I think it has been definitely felt by the cLabs Mento team, and also individuals in the Celo community, that being able to react and make certain parameter changes faster would have been desirable, and may still be needed. I’ve seen indications from other web3 projects that this is something they’re feeling too. In particular, I think this would be valuable around e.g. tools and parameters that exist in the protocol to bolster or manage the reserve.
I’d love input on three topics:
- speeding up regular governance votes
- thoughts on buy-in for broadening use of emergency hot fixes
- medium-term (weeks, months) changes we can make
Regular Governance Votes
All of this of course is a balance between inclusivity, security, and responsiveness.
Currently, a full governance proposal from on-chain proposal, takes 9 days from start to finish, not including CGP, prior discussion on the forum, or on calls etc:
- Dequeue (upvoting) is 1 day
- Approval (governance approver multisig) is 1 day
- Referendum (voting) is 7 days
Governance:
concurrentProposals: 3
dequeueFrequency: 1 day
minDeposit: 100000000000000000000 (~1.000e+20)
queueExpiry: 4 weeks
stageDurations:
Approval: 1 day
Referendum: 1 week
Execution: 3 days
The upvoting stage was designed as anti-spam mechanism and has not proven very necessary so far. We could reduce it to e.g 4 hrs.
Shortening Approval is hard because it currently requires 3 multisig approvers to sign, and the proposal fails if within 24hrs of the first signature, the remaining 2 have not signed.
The Referendum stage could be reduced, perhaps to 3 days, with good communication. There’s a danger that doing this might cause proposals to fail if they don’t reach the required participation level for the specific action in the proposal within that shorter time window.
Emergency Hot Fixes
We also have the emergency hot fix mechanism, designed to apply zero day fixes to Core Contracts without them sitting in public view for 9 days. Here, two thirds of validators can pre-approve a hash for a proposal so it can be executed immediately. This mechanism bypasses CELO voters, but could be used with rough consensus from validators in combination with an off-chain polling mechanism in wider circumstances.
Other ideas
Longer term, we can explore adding support for LockedGold and ReleaseGold to a tool like Snapshot.
We already discussed making the approval phase run in parallel with the referendum stage, and potentially removing the Dequeue phase.
We could also consider mechanisms for Governance to delegate control over certain parameters to other addresses, e.g to enable a multisig of reserve stewards.
Would love thoughts on this!