Project governance

peerchemist · November 18, 2020, 5:04pm

Still trying to decide about that, communication is going of for some time now.

The way I see it worst that can happen is that we design a good governance system in vain

Nagalim · November 18, 2020, 5:33pm

Then they don’t get to vote. I don’t think this is a concern, though I understand what you are implying about who constitutes the voter base.

The snapshot date should not be declared in advance. If individuals buy coins in anticipation of a vote without a clear announcement of a snapshot date, then I believe they are better than the traditional momentary pump and dump, and should be treated as having stake in the system for the purposes of the election. Because really we are trying to do a less-formal proof of stake here in the context of elections.

This is the theory anyway. I still think it would be good to give the foundation blackballing power.

peerchemist · November 18, 2020, 5:36pm

This leaves space for corruption in the org which organizes the elections. I still prefer that it’s per solved block in past year, at value of UTXO which has solved the block - cumulatively to the day when the ballots are sent out.

Nagalim · November 18, 2020, 5:56pm

What happens if you mint twice with the same utxo? Are there opportunities for gaming the system by using larger outputs? What if you mint once, move the coins, mint again at another address? The assumed solution to that is to use coindays instead. I also fear that this approach devalues continuous minting as opposed to those that mint one large output once a year (preferably just before the vote, if we’re using coindays). Also, it entirely cuts out traders from the process, though the cost/reward of that is definitely a philosophical argument.

I would prefer 1 mint = 1 vote or mint rewards = votes as opposed to this system.

peerchemist · November 18, 2020, 6:32pm

Gets summed up per this proposal.

Yeah, we may be forced to go with this.

zattack · November 18, 2020, 8:32pm

Let’s look at the best case scenario too: we put something simple in place quickly that can be improved upon and get to the same place but with money in the bank asap

Sentinelrv · November 18, 2020, 8:49pm

I just wanted to suggested something. Mind you I am not a coder and don’t know what’s involved or if it’s even possible.

Nu’s Motion Voting (1 vote per minted block)

Many of us here are familiar with NuBits and their voting system. In Nu’s motion voting, every time you mint a block, you add a vote to the chain. The more blocks you mint, the more votes you add. Those with large stakes get a bigger say than those with small stakes. The reason for this is because large stakeholders have more chances to mint blocks. This is completely fine in my opinion, as your voting weight should be proportional to how much stake you own.

In Nu’s model, the stakeholder can only add votes to the blockchain by minting blocks themselves. Unfortunately, this can take a long time, as you need to wait for enough blocks to be minted before you start seeing results. Nu has 1 minute blocks, while Peercoin has 10 minute blocks. If we were to use Nu’s voting system, it would take Peercoin 10 times longer to get results, because of the difference in block spacing. Small stakeholders may also not be able to get their vote counted due to the smaller chance of minting a block within the time window of voting.

Utilize Memory Pool for Quicker Voting

I was wondering if it’s possible to treat votes like we do transactions. Users would pay a transaction fee and cast their vote using the client. Once cast, the vote gets transmitted to Peercoin’s memory pool and immediately gets included in the next block, no waiting required. This prevents voters from needing to rely on the chances of minting their own blocks.

Instead, voters can take advantage of other people in the network who are minting blocks. This allows many people, even those with small stakes, to quickly get their votes included in the blockchain. They just pay the fee to send their vote out into the network, it gets collected by the memory pool while waiting to be included in the next block.

This would speed up voting considerably, compared with Nu’s model. For small stakeholders, instead of waiting possibly weeks or months to mint a block with your included vote, you could do it in about 10 minutes by relying on the next block producer to include it.

Rules & Conditions

Of course, it would require some extra rules to be programmed in for it to work right…

1. Every vote should include a weight by taking into consideration the size of a user’s stake.

2. Multiple votes for the same issue need to be rejected by the network. Only one vote should count.

3. Votes should only be accepted if the address you are voting with has minted a block within a certain time period. This proves you are an active participant of the network, and not simply a speculator.

4. Question to consider: Should your voting weight be dependent on your balance at the time you submit your vote, or what your balance was the last time you minted a block?

5. Question to consider: What happens if a large stakeholder has been sitting on their PPC for years without participating in security? Suddenly they see an election is happening, so they activate minting to produce one block just so they become eligible to vote. In my opinion this kind of stakeholder should not get a say, since they are not an active participant in the network’s security. Is there some rule that can be coded to identify lazy stakeholders like this so the system can reject their vote?

Criticism

Peerchemist told me the following…

I believe the 3rd rule above should help prevent this. Basically, all votes by an address would be rejected by the network if the voter has not minted a block from that address within a specified time period. It takes 30 days to become eligible to mint blocks. As long as you schedule the start and end of the election before 30 days, people buying PPC off exchanges should not be able to influence the vote, since it will end before they become eligible.

Again, I’m not a blockchain coder, so there is probably a lot of stuff I’m not considering. I just wanted to put it out there as an option.

peerchemist · November 18, 2020, 8:59pm

Missed this on my first read. This is good.

peerchemist · November 19, 2020, 1:55pm

Started writing some scripts to get understanding on what is to be done here.

One interesting info I got so far is that 1469 unique addresses have minted between block #450000 (2019-08-30) and block #530000 (2020-11-10).

jacob · November 19, 2020, 1:56pm

That is very interesting info.

Keep the stats coming!

jacob · November 19, 2020, 1:59pm

I’m really not sure about blackballing exchanges.

The fact of the matter is users have put their coin there, giving the exchange very significant responsibility. So another way of looking at this is simply setting the standard that all coins and all addresses are the same and if your coin is on an exchange, they are going to end up being able to vote with your stake.

Nagalim · November 19, 2020, 6:14pm

Using mint rewards for votes satisfies my concerns, including those implied by the blackballing concept. Mint rewards are in my opinion directly proportional to participation in the governance of the chain, to the best estimate we currently have.

backpacker69 · November 19, 2020, 6:28pm

i like that idea also, mint rewards become votes

peerchemist · November 21, 2020, 12:09pm

Here is a quick python script which finds voters and their vote count.

gist.github.com

https://gist.github.com/peerchemist/fda4b85fcccbb51a05187e46fbb86195

find_voters

from peercoin_rpc import Client

# block found 2019-10-27
starting_block = 460000

node = Client(username="rpc", password="pass")

votes = []

def get_block_minter(blocknum: int):

This file has been truncated. show original

peerchemist · November 21, 2020, 12:27pm

I guess it should be ported to js, so it runs in one of those online REPL consoles and can easily be ported to a web app later on.

Sentinelrv · November 23, 2020, 10:42am

@jacob linked this Twitter thread above. I thought it was pretty enlightening, so I just wanted to draw some more attention to it. You don’t have to read the whole thing, as most of the important points are laid out in the first half…

Elad Verbin:

DAOs & Token-based Governance is a fantasy.

Stop emphasizing token-based voting; start focusing on assigning executive power. Even countries and big corporations fail to ensure wide participation by their “token holders” - instead they maintain strong executive bodies

The crypto world has a fantasy about token-based governance. It goes like this: you want to have a commercial organization with a mission (e.g. a DAO). So you mint tokens that represent both shares in the profit and governance rights in the organization. Exactly like stock.

Now, the token holders are shareholders, so the theory is that they are incentivized to steer the organization to success. Just like in public companies! Shareholder governance, isn’t it great!

But the actual world doesn’t work this way.

In public companies in the real world, shareholder governance is not how decisions get made. The actual decisions are made by the executive bodies. The CEO, the senior leadership, the VPs, sometimes the board.

In public companies, shareholders are rarely called to make decisions, and when they are, it’s mostly to elect executives. Executives are important: they’re people in the know, with relevant skills, able to dedicate time and to wield power, and who are paid to do their job.

This pattern is pervasive. In democratic countries, public voting is held every several years, to elect members of parliament (~ board members) or to elect executives directly. In countries with strong direct voting, you might vote on approving policy, but never on setting it.

Executives are the ones setting policy, negotiating it, making day to day decisions. Their job is to be well-informed, to make data-based decisions, and to execute them. They are continually judged by the CEO, and the CEO is judged every several years by the token holders.

Big organizations – countries or corporations - always rely on the executives for all their operations, and only ask “token holders” for their input rarely, and on broad topics. Because token holders hold too little equity, they’re not incentivized to be well-informed.

So I don’t get why DAO advocates spend so much of their time obsessing about voting works and how token-holder value accrues, and so little time specifying how the executive works. Many DAOs don’t even specify an executive – they pretend someone will take the mantle organically.

Designers of DAOs and other token-voting schemes need to focus much more on designing the executive. For each 100 words you spend on describing how the executive works, you’re allowed to spend 20 words on specifying the token holders dynamics, processes and incentive design.

If we keep pretending that token-based governance is a thing, we’ll just keep designing dysfunctional systems that only serve as fodder for speculators. If you want to design tokenized systems that thrive, study how real-world companies work: they rely on strong executive bodies.

Just to add: my point is not that real-world companies or state democracies have great governance. My point is that even in these big important systems, the “token holders” have little interest to participate beyond voting once every few years – so surely your rinkydink DAO shouldn’t expect thousands of tiny token holders to make executive decisions. As @CemFDagdelen points out, corporate governance and state democracies are not the end of history. But the vector for improving them is not the vanilla “more direct democracy”

Direct Democracy pretty much does work as-is. For many reasons, but primarily because of the paradox of voting: small stakeholders are not incentivized to put in the work to make good decisions. token-based governance has offered few ways to solve this.

Also, to be clear, there are some promising suggestions for token-based governance that goes beyond the “direct token-based voting” paradigm. E.g. @joincolony @DemocracyEarth @daostack @AragonProject and others. But most projects are naive…

I see way too many projects and whitepapers and discussions that just assume as an axiom that vanilla token-based voting is a decent solution for governance. Instead, they should say “this is a hard problem, we’ll use best-of-class solutions as they develop”.

Another point: you can try to design DAOs that don’t have empowered executive bodies, but do have sophisticated voting systems that try to solve the paradox of voting. (cf. Liquid Democracy or Holographic Consensus.)
I claim this is insufficient and that executives are necessary:

Executives of a body can negotiate. They can set long-term plans and then adapt them on the fly. They are empowered to represent the organization. And they have this power in a legally-binding way:
Politicians sign peace treaties before bringing them to congress.

CEOs make offers to buy and sell companies, before bringing them to the board or shareholders (if at all). Executives wield dramatic power. They can be reprimanded for their actions – but only after the fact. Empowered executives of this sort are necessary for organizations:

If your DAO only makes decisions by voting-based governance, and if potential decisions are only “brought up for voting” rather than executed, then there are many modes of organizational action that are completely blocked to you:

Outsiders cannot negotiate with an organization that does not have empowered executives that are allowed to represent it and make decisions for it. Iterating on policies becomes impossible. It’s like trying to run a country with a parliament without a government.

My rule of thumb: if your token governance design, or your DAO, does not have executives, i.e. people who can take binding actions first, and only be punished after the fact, then you’ve designed a dysfunctional governance system.

One good example for a system with empowered executives is @joincolony, which is designed to allow people to make empowered binding decisions first, and then to later suffer the consequences if they angered the group. I wish this concept was more widely used.

Here’s a really important caveat to my claims, by @VitalikButerin. He says that DAOs have many roles. Some DAOs are “wide DAOs” that have wide responsibilities. Some DAOs are “narrow DAOs”, entrusted to manage one narrow system semi-automatically

He says that on narrow DAOs, you absolutely don’t want empowered executives – you want a stable and dependent organization, that just quietly does its job. Like the Bitcoin network, for example. He references the work of Thomas Sowell on this dichotomy.

Vitalik Buterin:

Thomas Sowell talks about this a lot: there is an inherent tradeoff between flexibility/agility on one side, and stability/dependability on the other side. Different systems benefit from being at different points in this tradeoff space.

There’s a big spectrum of DAOs’ missions, and they need very different designs. In fact, I already have a diagram for this, from my blog post on behavioral crypto-economics. The more you are on the top-left, the more you’d want an empowered executive.

Also: DAOs today already do have executives. They’re just self-appointed, unofficial, de-facto executives, but no duty of fairness, nor checks&balances, nor the right to officially represent the org.
Today’s DAOs are not free of executives, they just have dysfunctional executives

You can speculate that this trend will get fixed with time by better governance methods, eliminating de-facto executives. But I suspect executives are here to stay, and you must choose to either design them into the system, or to deal with dysfunctional de-facto executive bodies.

Great thread from @EAThomson pointing out some typical weaknesses of executive bodies:
https://twitter.com/EAThomson/status/1316315253425221635
To be clear, corporate executive bodies can serve as inspiration, but we should do even better!

And a good point from @Sam_Liban about the importance of intransparency:

Sam Liban:

So true & well put.
I would add, that also few talk/think about the levels of intransparency organizations need to actually function…

If executives & all others would be 100% transparent to investors, there would be no room for a pivot or any creativity beyond initial ideas.

To see why you need to design executive power into your DAOs, just look at thepower-jockeying that happens when a DAO is lacking a defined executive. In the end, a bunch of players all try to grab power, and often the worst players, or the ones with foreign interests, win.

Recall that a key feature of a Democracy is the peaceful transfer of power. When you avoid defining who has power and how it’s reassigned, you don’t get a system where no one has power, but a system where the worst entities seize power and monopolize it.

peerchemist · November 23, 2020, 12:36pm

Very good post.

RobertLloyd · November 23, 2020, 5:30pm

I totally agree with the article. I’ve been on a Committee for 20 years (for an organisation of 300 people) and all the decisions and activity are done by a small number of people (five at most).

We used to have Annual General Meetings, when members voted for committee officers, but usually only a tenth of ordinary members would attend, and the votes were really just a formality. The fact is that most members are happy to let officers “get on with it”, and so long as members get updates, they keep rejoining.

We eventually abandoned the AGM, as it did not resolve problems. We found that it is more important to have controls within the committee, such as a required level of transparency, and ensuring a second person has access to passwords, the membership list, etc. in case an executive officer goes awol.

peerchemist · November 23, 2020, 5:57pm

How I imagine the elections to look like / be organized:

Glossary:

Foundation - Stichting Peercoin Foundation, a legal organization established to serve as legal face of the Peercoin project.
Voters - minters who have solved a block at least once in last year.
Leader - leader of the Peercoin project.
Team - cabinet led by the Leader.

Timeline of elections

Foundation proclaims that there will be the elections and explains the rules to community.
- Rules are the threshold, the duration of the elections and other details like that.
Foundation distributes the ballots to voters.
Voters deposit ballots in the address which matches their favorite candidate.
Upon crossing the threshold of support for the candidate (66%+ per current proposal), Foundation proclaims the winner and starts the transfer of control to the elected Leader and the Team.

Technical details

It is best all of this is made into a webapp and as it’s the norm for all our present webapps hosted on the GitHub so clients load it directly from the github. No servers involved. There is no need to communicate with any external servers as all can be deduced and calculated from the chain directly.

For example: governance.peercoin.net

A sidenote:

I had someone ask me what is a webapp, so I’ll explain. It’s a website. Yep, just a website. Website which can do all the features explained in this post.

Ballots

Ballots are found by a simple algo:

Find all PoS blocks in the past year
Find out what was the mint reward (block subsidy) of all those blocks
Find the pubkey which found the block (address)
If pubkey has found a block more than once in the past year, reward is summed up.
- Sum of all the mint rewards is referred to as “weight” later in the text.

You can check the sample code here.

Ballot is a min-value UTXO, think 0.01 PPC or so. It is sent manually from the Foundation to each eligible pubkey. Foundation is sending the ballots out to not disrupt the minting process of active minters. They do not have to burn coinage by moving coins around.

Votes

Votes are ballots weighted by stake.
Each candidate is associated to an address. Votes are cast by simply transferring the ballot into the address which matches the voters favorite candidate.

User perspective

How I imagine voting would look like from the perspective of the user.
All info would be present on the governance.peercoin.net, voters, ballots, vote counting and the process of voting itself. Everyone would be able to validate data on their own, as there is no central server it’s all parsed from the blockchain directly.

Active minters would go to governance.peercoin.net and run their local Peercoin node with allowed RPC connection. The webapp would query the node for ballots, and if there are ballots received it would prepare the voting transaction. Such transaction would be loaded into the Peercoin wallet, signed and sent.

That’s it.

Results of voting could be followed in real time on governance.peercoin.net.

Nagalim · November 23, 2020, 9:18pm

This is a good description of a basic strategy for voting. I would like to point out one issue I have, which is with reference to the 66% supermajority threshold. If we choose to use anything other than a plurality, we need to think about what happens if no candidate achieves the threshold required. Having to do the vote multiple times is inefficient, as is advanced voting options such as 'ranked choices, which are known solutions to avoid developing a 2 party system which plagues plurality-based voting. I know you expect somewhat unanimous voting, but why have voting at all if we can’t accommodate the possibility of a close race between two candidates? If we have two really good candidates, and one gets 51% while the other gets 49%, what happens? It doesn’t make sense to do nothing and keep the old leader, who may be neither of these candidates. Supermajority voting doesn’t really work for elections where a decision must be made and there is no default ‘do nothing’ option.

One option for a form of ranked choice would be:
3 candidates, 2 of which are similar and one very different. The two that are similar make it clear that when one of them receives less votes than the other, they will give their votes to the other candidate. In this way, the two addresses no longer represent just 2 separate candidates, but rather they represent ranked choice votes with one candidate first and the other second. We could hypothetically do full ranked choice this with any number of candidates, though the number of addresses grows quickly with the number of candidates. 2 candidates, 2 addresses. 3 candidates, 6 addresses. 4 candidates, 24 addresses.