Here @sumoshi21 makes the KISS argument. Keep it Simple Stupid.
And I think it’s a very reasonable posture to take.
He argues to limit the scope of governance to explicit governance methods which would remove the need to consider value transfer.
However, simplicity cannot outweigh network security.
I asked earlier if it makes sense for all governance actions to be considered equally.
My opinion is no. We will have Type 1 vs Type 2 decisions. Some changes like sporks are existential and can halt or fork the network, while some changes in network variables only change relative winners and losers. Does it make sense to have the same quorum requirements and voting periods for both kinds of governance decisions? My opinion is no.
Regarding 51%, one comment sumoshi makes is:
If you can’t count on them to vote, then you can’t count on them to upgrade the node, and then we’ve got bigger problems
Robust protocol design means anticipating worst case problems. Yes it’s a big problem, but we have to provide some sort of answer.
Let’s consider Bitcoin. Bitcoin PoW is open and decentralized because if a portion of miners drop out, the protocol continues to function just fine with the existing hashrate (albeit with less security), and new hashrate can come in from anywhere. PoS is not the same.
@sugoibtc brings up concerns that 51% may be too high of a threshold. And for Type 2 decisions or rubber stamping that doesn’t materially affect individual pillars, it may be a valid concern.
For Type 1 decisions, I actually have the opposite concern. That 51% may be too low. Many PoS consensus mechanisms networks are only provably (mathematically) secure assuming a supermajority of honest actors or 67%. No one has done such rigorous analysis on Zenon but supermajorities are often the standard.
Politically when it comes to big changes, 51% has a big chance to fork the network. It is not a resounding consensus. Will pillars who lose 49% to 51% choose to just lay down? Especially if delegators and other network participants agree with them? It is very plausible that they choose to fork and then win in the long term. It’s much less plausible with a supermajority.
(We haven’t even gotten into governance which includes other actors, something many community members want considered. I personally think for network security that hard binding votes should be restricted to pillars due to the burn requirements and skin in the game.)
One possible solution is to start with a generic action method like sumoshi has but to give it a dynamic quorum requirement. It starts with supermajority quorum and YES vote requirement with a relatively long duration, but after a certain point if it doesn’t meet quorum then every 2 weeks, quorum drops by a certain amount/percentage. This would allow Zenon to be robust and survive without a hardfork if we lose significant pillars. This would be the path for Type 1 decisions and the the default governance method.
For Type 2 decisions, perhaps it would make sense to implement other methods in the future that are limited in governance scope with less stringent quorum requirements. Basically a whitelist of safer governance decisions.
One concrete suggestion on sumoshi’s proposal is that perhaps it makes sense to keep the 1 ZNN requirement but also require the vote be initiated by a pillar.