译文：区块链治理说明

本文由“币嗨Bihi内容合伙人计划”赞助

注：这是V神博客中的一篇文章《Notes on Blockchain Governance》的第三部分

Low VoterParticipation

低迷的投票参与度

One of the main criticisms of coinvoting mechanisms so far is that, no matter where they are tried, they tend tohave very low voter participation. The DAO Carbonvote only had a voterparticipation rate of 4.5%:

迄今为止，货币投票机制的主要批评之一就是，无论他们在哪方面努力，持币者的参与度都很低。 DAO Carbonvote只有4.5％的投票参与率：

Additionally, wealth distribution isvery unequal, and the results of these two factors together are best describedby this image created by a critic of the DAO fork:

此外，财富分配非常不平等。DAO分叉的批评者，他们创作的图像很好地展示了这两方面因素的影响

The EIP 186 Carbonvote had ~2.7 million

ETH voting. The DAO proposal votes did not fare better, with participation never reaching 10%. And outside of Ethereum things

are not sunny either; even in Bitshares, a system where the core social

contract is designed around voting, the top delegate in an approval vote only

got 17% of the vote, and in Lisk it

got up to 30%, though as we will discuss later thesesystems have other problems of their own.

EIP 186 Carbonvote拥有约270万次的以太坊投票。DAO提案的票数并不好，参与率从未达到10％。除了以太坊，其它的事情也是阴云重重;在Bitshares这个系统中，核心得社会契约是围绕着投票设计的，最高代表的赞成票只有总票数的17％，在Lisk系统中，最高代表的赞成票会上升至30％。尽管这些系统有自己的问题，我们稍后会再讨论。

Low voter participation means twothings. First, the vote has a harder time achieving a perception of legitimacy,because it only reflects the views of a small percentage of people. Second, anattacker with only a small percentage of all coins can sway the vote. Theseproblems exist regardless of whether the vote is tightly coupled or looselycoupled.

低参与度意味着两件事。首先，投票难以实现合法性，因为它只反映了一小部分人的观点。其次，拥有一小部分货币的攻击者就可以左右投票结果。无论投票是紧密耦合还是松散耦合，问题始终存在。

Game-TheoreticAttacks

博弈论的攻击

Aside from “the big h5ack” that received

the bulk of the media attention, the DAO also had a number of much smaller

game-theoretic vulnerabilities; this article from HackingDistributeddoes a good job of summarizing them. But this is only the tip of theiceberg. Even if all of the finer details of a voting mechanism are implementedcorrectly, voting mechanisms in general have a large flaw: in any vote, theprobability that any given voter will have an impact on the result is tiny, andso the personal incentive that each voter has to vote correctly is almostinsignificant. And if each person’s size of the stake is small, their incentiveto vote correctly is insignificantsquared. Hence, a relativelysmall bribe spread out across the participants may suffice to sway theirdecision, possibly in a way that they collectively might quite disapprove of.

除了受到大部分媒体关注的“大黑客”之外，DAO还有一些小得多的博弈论漏洞; 这篇来自Hacking Distributed的文章做了很好的总结。但这只是冰山一角。即使投票机制的所有细节都得到了正确实施，投票机制通常也存在很大缺陷：在任何投票中，特定投票人对结果产生影响的可能性都很小。因此，让每个选民都正确投票几乎是不值一提的。如果每个人的权重都很小，他们正确投票的动机是微不足道的。因此，在参与者身上散布的相对较小的贿赂可能足以影响他们的决定，可能以他们完全不同意的方式行事。

Now you might say, people are not evilselfish profit-maximizers that will accept a $0.5 bribe to vote to give twentymillion dollars to Josh arza just because the above calculation says theirindividual chance of affecting anything is tiny; rather, they wouldaltruistically refuse to do something that evil. There are two responses tothis criticism.

现在你可能会说，人们不是邪恶的自私利益最大化者，他们会接受0.5美元的贿赂投票给Josh arza两千万美元，因为上面的计算表明他们影响任何东西的个人机会很小;相反，他们会无私地拒绝去做邪恶的事情。对于你的这种批评，可从两方面回应。

First, there are ways to make a “bribe”that are quite plausible; for example, an exchange can offer interest rates fordeposits (or, even more ambiguously, use the exchange’s own money to build agreat interface and features), with the exchange operator using the largequantity of deposits to vote as they wish. Exchanges profit from chaos, sotheir incentives are clearly quite misaligned with usersandcoinholders.

首先，制造一个相当合理的“贿赂”有很多种方法; 例如通过提供存款利率，交易所运营商可以使用大量存款按其意愿进行投票（或者，交易所偷偷摸摸地使用自己的资金来建立良好的交易界面和特征）。交易所能从混乱的局面中获益，所以它们的动机明显与用户和货币持有人不符。

Second, and more damningly, in practiceit seems like people, at least in their capacity as crypto token holders,areprofitmaximizers, and seem to see nothing evil or selfish about taking a bribe ortwo. As “Exhibit A”, we can look at the situation with Lisk, where the delegatepool seems to have been successfully captured by two major “political parties”that explicitly bribe coin holders to vote for them, and also require eachmember in the pool to vote for all the others.

其次，也更令人沮丧的是，在实践中，人们似乎至少以加密代币持有者的身份，是利润最大化者，在接受一两次贿赂的时候不会意识到邪恶或自私的地方或自私行为。

也更有说服力的，那些利益最大化的人，在作为token持有者的能力范围内，在实践中人们似乎至少以密码令牌持有者的身份是利润最大化者，并且似乎没有看到任何贿赂或贿赂。在“图表A”中，我们可以看看Lisk的情况，代表团似乎已经成功地被贿赂投币者投票的两个主要“政党”所俘获，并且还要求投票池中的每个成员为其他所有人投票。

Here’s LiskElite, with 55 members (outof a total 101):

这是LiskElite网站的55名会员的投票情况（共101名会员）

Here’s LiskGDT, with 33 members:

这是由33名会员的LiskGDT网站

And as “Exhibit B” some voter bribes

being paid out in Ark:

图表B中显示了一些投票者接受ARK贿赂的情况

注：ARK是一种代币。

Here, note that there is a keydifference between tightly coupled and loosely coupled votes. In a looselycoupled vote, direct or indirect vote bribing is also possible, but if thecommunity agrees that some given proposal or set of votes constitutes agame-theoretic attack, they can simply socially agree to ignore it. And in factthis has kind of already happened - the Carbonvote contains a blacklist ofaddresses corresponding to known exchange addresses, and votes from theseaddresses are not counted. In a tightly coupled vote, there is no way to createsuch a blacklist at protocol level, because agreeing who is part of theblacklist isitselfa blockchain governance decision. Butsince the blacklist is part of a community-created voting tool that onlyindirectly influences protocol changes, voting tools that contain badblacklists can simply be rejected by the community.

在这里，请注意紧密耦合和松散耦合投票之间存在一个关键性区别。在松散耦合的投票中，直接或间接投票贿赂也是可能的，但如果社群同意某些提议或一组投票构成博弈论的攻击，他们可以简单地同意忽略它。事实上，这种情况已经发生过 - Carbonvote包含一个与已知交换地址相对应的地址黑名单，并且这些地址的投票数不计算在内。在紧密的投票中，无法在协议级别创建这样的黑名单，因为同意谁是黑名单的一部分就是区块链治理决策的本身。 但由于黑名单是社区创建的投票工具的一部分，只会间接影响协议变更，因此包含黑名单的投票工具可能会被社群拒绝。

It’s worth noting that thissectionis nota prediction that all tightly coupled votingsystems will quickly succumb to bribe attacks. It’s entirely possible that manywill survive for one simple reason: all of these projects have founders orfoundations with large premines, and these act as large centralized actors thatare interested in their platforms’ success that are not vulnerable to bribes,and hold enough coins to outweigh most bribe attacks. However, this kind ofcentralized trust model, while arguably useful in some contexts in a project’searly stages, is clearly one that is not sustainable in the long term.

值得注意的是，这一部分并不是一个关于所有紧密耦合的投票系统都会很快屈服于贿赂攻击的预测。完全有可能有很多人会因为一个简单的原因而生存下去：所有这些项目都有手握巨资的创始人或基金会，他们充当大型集中演员，对平台的成功感兴趣。这些人并不容易被贿赂，并且拥有足够的货币去打败大多数贿赂攻击。然而，尽管在项目早期阶段的某些环境中，这种集中式的信任模式很有用，但它显然是不可长期持续的。