Thoughts on Governance

May 15, 2019

This week’s CW we got a speaker Jian from EthFans talking about the governance mechanisms in blockchain systems.

This is the second time for Jian to talk about this. There’s a recap online for his last talk on this.

Here are some points we have been talked about.

On-China Governance.

Pure on-Chain governance is a bad idea. The idea is that basically instead of letting the block producers to determine what should be the state stored on-chain and how should the chain works, on-chain governance mechanism (e.g. Tezos) suggests that we should have another voting mechanism implemented on-chain that let people vote for the governance result with their voting tokens. This is a bad idea as it tends to take the governance power away from the block producers (BP). BPs are competitors to each other as they want to fight for the right of producing blocks, so it is nature and healthy for them to vote for the governance of the blockchain system. However, if we have another mechanism that give the power of voting to someone else, these may not necessarily have the competitor relationship to each other.

What is the Goal for a Governance Mechanism?

So a lot of people are talking about it is very important for public blockchain to have a governance mechanism, but not a lot have explained why. It would make sense if we want to design something to solve a specific problem. However when it comes to governance, I think we have no idea what is the problem we are trying to solve here.

Some people may argue that fork is bad, and we should use governance to avoid fork. Some may think that it’s too hard to update the public blockchain system, so we need an explicit flow for making changes to the blockchain system.

The problem here is that bad and good are always a matter of perspective. When a fork happens, it definitely mean that someone is gonna be happy for this fork to happen. Because if no one is happy about doing a fork, then there will be no fork in the first place. Same to the upgrading a public blockchain system.

But making people happy is still a good starting point for thinking about what is the objective of a governance mechanism. The existence of government is also for making every person governed by them happy. At least this is what a government may claim.

So if we make the objective of designing a governance mechanism be making every user of a public blockchain happy. Then maybe we can refer to A Theory of Justice from John Rawls. This basic idea in it is that to bring justice to the society, we should emphasize on freedom of choice. This means that everyone should have the freedom to make their own choice.

I personally agree with Rawls on this. In a public blockchain system, we should let every user to have the freedom to make choice for every transactions or even the future development direction of this public blockchain system.

Although in this process, some people may have more vote than others because they have more accumulated capitals that can be converted into votes, this should not obey the theory of justice. Some may have more willingness to vote, and some may just don’t give a shit about how this system works.

According to the theory of justice, all of these does not matter because all of them have been given the freedom of choice. So even if some of them may still complaint about the final result, there’s nothing more we can optimize here.

Let Great People have More Votes

Someone from the discussion mentioned an idea of making a reputation system to let better people have more votes for governance. I personally think this is neither necessary nor easy to do.

In my argument, people that have some good attributes or properties should naturally result in a larger position of capitals (averagely speaking). The capital here could be money, or it could be influence.

These “great” people can then use these capitals to “buy” more votes in a governance. It means that they already have more rights for voting than others. So I don’t think there is a need to give them another layer of weighting, according to the same properties and attributes.

It is also hard to build such a system. If this system has some rules defined by humans, then some humans could definitely figure out the rules and optimize themselves according to it.

So AI based reputation systems here would be an appealing solution. But I don’t know how would that goes.