译文/Translated：
最近，关于EOS上的超级节点(BP)是否应该遵循基于EOS宪法产生的争议解决仲裁有很多争论。关于为什么你应该希望你的区块生产者遵循这样的仲裁，理由其实不胜枚举。
1/ 作为诚信的标志
遵循裁决是BP展示它们支持整个社区以及社区行为合理的载体—EOS宪法。本质上来说，EOS宪法是和整个社区的合约，也是约束整个社区的合约，展示你遵循合约对于长期的信赖关系是非常重要的：言行合一。
如果BP不执行根据EOS宪法第九条产生的裁决，它们实际上是在挑战EOS宪法。尤其是，EOS宪法第九条明确表示所有的争议都应该用这种方式解决（其中包括很多制约和平衡）。这本是一条强硬的条款，在一般的合约中，如果违反了这一条款，那么整个合约都危险了。
现在对于用户的问题就是，如果你的BP有不遵守合约的恶名，你还希望你的区块产自它吗？
2/平衡
管理安排和EOS宪法达成了平衡。合约的立场非常清晰：EOS宪法设定所有规则的高层框架，但因为争议总是存在的，这时候第九条EOS宪法就描述了如何解决冲突。这一条款适用于任何人，除非你已经签署了EULA（作为dapp端）或做了其它安排从而退出合约。
EOS宪法创造了强大了三角凳——BP<->持币人投票<->争议解决。我们必须保持社区的这个结构，因为三角形具有稳定性。三角凳一条腿断了，其它两条也会受到影响——由此带来的连锁相应会削弱整个区块链。倘若宪法被打破，那么EOS就会走上Steemit的老路。
3/ 平等
如果BP放弃了宪法，但是其它成员却受到它的约束，那又产生了一个问题。一个很大的哲学问题是，BP算成员之一吗？如果一群人因为某些裁决或者合约让他们感到不舒服他们就拒绝接受裁决或撕毁这些合约，那我们真的希望这样一群人插手我们的区块？但如果我们接受区块生产外包的话，那我们也可以通过招标、需求建议书、或者直接去找优秀的区块生产企业（Strongblock? Consensys? R3？等）和他们签订合同来达成目的。
至少这样还会让社区拥有合约的确定性——所有企业都知道，尽管这不像加密的确定性那样吸引眼球，但这种确定性是必不可少的。既然EOS是在做生意，那像“代码即法律”这样打马虎眼的协议就必然不是我们的诉求。做生意总是需要某些形式上的确定性。撕毁合约就赤裸裸地证明我们在BP上没有什么确定性。
还是说BP是某种更高的存在，甚至高于我们都坚守的宪法的存在？那我们是搞了一个精英和外来者的社区吗？是不是只要有人说自己是BP他就可以立于宪法之外了？这一切的终点会在哪里？
4/ 保护成员
区块链中能出现争端解决方案是具有革命性的。它的存在包含两层含义：首先它能解决争端，所有的社区中都存在争端。我们经常在探讨DAO这个经典案例，但这样的例子还有很多，这些例子中区块链都把宝贵的编码和做生意的生意浪费在争端上了，有些因此走上了岔路，有些给商业和贸易带来了灭顶之灾。区块链中已经损失了几十个亿了，我们希望在我们的区块链中能减少一些损失。
第二个理由是保护成员——成员的人身财产安全。我们希望能够保护成员，因为只有商业环境安全，它才有利可图。
否则，最终的结果不过是野蛮生长的新行业。但没有人会希望把他们的商业伙伴带到这个外表光鲜亮丽的区块链中，结果却惊闻他们要么被黑客袭击，要么是代币被送到什么不存在的链上，要么是被一些诸如multisigzykthingumybob之类看上去牛逼哄哄的账号劫持账号被锁定，要么就是虽然自己严格按照指示操作代币还是全部丢了。
5/ 减少BP滥用职权
因为BP有权执行非代码交易，那它们可能就真的会这么做。没钱了？没问题，我们可以从闲置账户中先挪一些出来。有些Dapp占用太多交易太烦人了？冻结他的账户。BP排名摇摇欲坠？挤到前一百就行了。没办法执行宪法？搞寡头把收益不高但表现良好的BP赶走就行了。服务Dapp赚钱？先安排适合自家伙伴的，帮他们赶走对手。交一点钱就行了。
历史证明BP总是抱团的。他们会先控制链，然后找一些特殊的方式赶走成本低廉但有创意的对手，最后给自己找到商业贿赂的来源。这不是什么稀奇的事情—实际上这种事情太普遍了—人性总是如此。
EOS的设计是要提供一系列的制约和平衡。以前的很多版本即出现了权力滥用的问题。但是争议解决方法提供了新的制约和平衡——BP们再也不能不承担任何风险就随意藐视规则。
但BP怎么会希望这么被“控制”呢？原因就比较复杂了——如果系统中有一个滥用权力的BP，诚实的BP现在面临两个选择：控制和赶走那个BP，或者同流合污。但同流合污的代价更低、收益更高、所以长远来开，如果没有任何制约的话，胜出的总是滥用职权那一个。DPOS特别容易受到寡头的影响——如果7个滥用职权的BP同意的话，他们可以组织任何投票，15个BP的寡头集团就可以通过任何投票。
这就是劣币驱逐良币定律，坏的BP最终挤走了好的BP。诚实的BP希望手上能有武器以免出现这样的问题。诚实的、心系社区的BP因此就更加希望制约和平衡，希望能够滥用职权，因为如果没有这些武器，他们知道未来会有多艰难。
6/ 保护BP
所有的成员或多或少都会受到很多攻击。审查、资产盗窃、监视、提取、勒索等，除此以外还有很长的威胁名单。对于BP来说尤其如此，它们都是好识别、容易遭到威胁的公司。威胁主要来自两方面——内部（上方）的和外部的。
外在的威胁来自于拥有更多资产和资源的大机构，它们可以为所欲为。外在机构可以强迫本地BP按照它们的路线发展，而不是沿着社区的路线。这样的例子太多了。这样的机构大起来可能像国家一样庞大，中等的也可能像黑社会，就算小的也可能是勒索犯或抢劫犯。在很多国家中，犯罪集团猖獗，拥有资产的人都受到威胁。暴力、绑架、谋杀这样的事情可谓稀松平常。
DPOS却给BP提供了避免攻击的武器——投票把它们赶出去。实际操作中，EOS还不能做到这点：正如Thomas Cox说的，（这里）没有积极的持币人。很多重要的持币人没有代表一些重要公司，诸如那些拥有高运营资产却没有太多储存EOS的。资深的内部人员往往拥有过多信息却没有太多EOS动摇投票结果。外在市场的信息薄弱、混乱、也容易被破坏。总之，每个故事之后都存在阴暗面。
争议解决给这些问题提供了独立的视角，通过裁决，它给社区传递了明确的信号。这样，它就和持币的投票人密切合作——通过传达可靠的信息，争议解决加强了DPOS反馈过程。妥善形成的争议解决可以快速地提供解决方案，因此让BP面对任人唯亲或者披着羊皮的狼的时候，它们都能够安然渡过。
争议解决机制让BP不能再通过代码不接受的方式对区块链做出改变。换言之，BP可以直接说，“我们没有权限，你提起争议解决。”Lost Key案件就是这么做得——BP正确地行使了宪法给他们的权力，提请争议解决论坛解决问题。
合同中不给予这样的权力就是一个反抗内部和外部权贵的有力武器。再加上选民的废除权，这两个武器可以很好地合作。
7/减少工作负担和认知压力
不管怎样，代码出问题的时候，BP总要做一些实事。但因为他们只是普通人，他们在做的时候会不断遇到压力，也总会有亲朋好友想让他们开个后门。然而，他们没时间也没耐心解决这么多需求。最后的结果也只会有这么两个：
- 只有在BP的好朋友们遭遇巨大的价值灾难时，紧急恢复机制才被启用，以太坊就是这样的。或者，
- BP（重新）发明争议解决机制，使其成为独立的决策部门，从而减少自己的工作压力，也让干预系统更加公平。
BP从根本上就不愿意决定关于权力的问题。这些问题实在太难了。争论一个密钥丢失案件是不是可以被解决这是一个社区问题，而不是一个技术层面的区块生产问题。EOS的设计去除了这样高层次的权力问题，一个原因是要求BP既要生产区块，又要解释社区权力，还要在思考过程中完全抛开自己的利益，还要让他们不受干扰，要依赖BP完成这么多任务实在太过分了。
有没有什么情况下不是这样的呢？
EOS固有的分离机制的好处我们已经提了很多了。但是没有什么计划可以一直走下去。什么样的情况下会出问题呢？
往小里看是裁决出错。往大里看是争议解决论坛可能反对社区利益。但您可以参考关于制约和平衡的文章，看该系统会如何避免犯这样的错误。
我们现在所处的刚好是第七点提到的：
7/ 进一步控制– 各方可能拒绝遵守裁决，这种情况下它们都违约了。接下来会遇到什么事不好说，毕竟这要看情况，但我们假设裁决存在严重错误。
这时，BP就会拒绝执行裁决。这是最严重的！因为他们是同侪之首——它们是执行代码、宪法、以及其它社区良好的机制的群体。如果BP因为愤怒就朝干瞪眼，论坛反过来也只朝着BP瞪眼，结果就是停滞状态——权力双方彼此不服，我们就陷入宪法危机。这最终还是社区要解决的问题。
因此，EOS就被施加了古老的中国咒语（*）—生活在有趣的时代。
（*）尽管英语世界说这是中国咒语，但实际上中国历史上却明显没有这样的话。
原文/Original：
There is much debate at the moment as to whether BPs in the EOS mainnet should follow dispute resolution rulings based on the Constitution. Here’s a non-exhaustive list of the reasons why you want your block producer to follow rulings from dispute resolution.
1/ As a signal of good faith
Following rulings is BPs’ chance to show agreement with the Community and its vehicle of reasonable behaviour – the Constitution. In essence, the Constitution is the contract with and for all the Community, and showing that you stick to your contract is important for long term trust: say what you do and do what you say.
If BPs do not enforce rulings issued under Article IX, then they are challenging the Constitution. Specifically, Article IX says that our disputes are resolved this way (with many checks & balances ). This is such a strong clause, in the general scheme of contracts, that breaking this clause means the entire contract is in danger.
The question for users then is, do you want BPs producing your blocks when they have a record of not sticking to the contract?
2/ Balance
The governance arrangement is balanced by the Constitution. The contractual position is clear: the Constitution sets the high level framework for all rules, and because there are always disputes, Article IX describes how to resolve disputes. It applies to everyone, unless you and your parties have contracted out with an EULA (as a dapp) or other arrangement.
The Constitution creates a strong three legged stool – BPs <-> Stakeholder voting <-> dispute resolution. It is important to keep this structure of the community, because a three legged stool does not wobble. Breaking one leg affects others – the domino effect causes consequences that weaken the overall blockchain. Breaking the Constitution reverts EOS back to Steemit.
3/ Equality
If BPs have abandoned the Constitution, but all others are bound by it, this creates a difficulty. One philosophical question to be posed is, are the BPs members? Do we want a group of people doing our blocks that deny rulings or tear up contracts when it is inconvenient to them? If we had wanted to outsource our block production, we could have put out a tender or RFP or gone to one of those fine enterprise blockchain providers (Strongblock? Consensys? R3?) and signed a contract with them.
That would at least have given the Community contractual certainty – something that all businesses know as essential albeit not quite as shiny and attractive as cryptographic certainty. As EOS is about business, we can’t just rely on handwavy arguments like code is law, businesses do actually need certainty of some form. Breaking contract is about as elegant a proof that we have no certainty in the BPs as it can get.
Or are BPs some superior form of being, above the Constitution that we all are stuck with? Are we building a society of elites and outsiders? Can anybody just step out of the Constitution just by declaring oneself a BP? Where does it end?
4/ To protect members
The presence of dispute resolution in a blockchain is revolutionary. It is there for two reasons: one is to resolve disputes, which always happen in all communities. We talk about the canonical case of The DAO but there are many others, all of them taking valuable time away from coding and business, and some of them causing forks or other calamities to business and trade. Billions have been lost in blockchain, we want to reduce that number some in our blockchain.
The second is to protect the members – themselves and their assets. We want to protect the members because profitable business comes to those who have a safer environment.
Then there’s the wild west. Nobody wants to bring in their business companions to this new shiny blockchain, only to hear they’ve been hacked, or they sent their tokens to /chain/null, or their bright shiny new multisigzykthingumybob snarked on them and locked it all up, or they followed the instructions and still lost it all.
5/ To reduce BP abuse
Because BPs can execute non-code transactions, they will be tempted to do that. Short of money? No problem, let’s just move some from somewhere where we know it is idle. Annoyed by some Dapp that is using too many transactions? Put a freeze on his account. Squeaky wheel BP? Let’s knock it back to the top 100. No way to enforce the Constitution? Let’s cartelise and block out the efficient cheap BPs. Money to be made servicing Dapps? Let’s arrange matters to suit our friends, and help them knock out their competitors. For a price.
History shows that the BPs will group together. They will start to control the chain, and find extra special ways to block low-cost innovative competitors, and to make a little more payola for themselves. It’s not special – indeed it is very ordinary – humans always do this.
The design of EOS is to provide a number of checks and balances. Earlier versions suffered many abuses. Dispute resolution provides a new check & balance – BPs can no longer flout the rules without some risk of consequences.
But why would BPs want to be ‘controlled’ in this way? The reason is a bit complicated – if there is an abusive BP in the system, an honest BP has two choices: To control and eject that BP, or to become as abusive as that BP. Abuse leads to lower costs, higher profits, so the abusive BP will always win out in the long run, if uncontrolled. DPOS is especially vulnerable to cartels – if 7 abusive BPs agree, they can block any vote, a cartel of 15 can pass any vote.
To echo Gresham’s Law, Bad BPs squeeze out the good BPs. Honest BPs want tools at their hands that make that squeeze harder. Honest, community-minded BPs therefore want more checks and balances. More controls on abuse, because without these tools, they know the future.
6/ To protect BPs
All members are more or less vulnerable to many attacks. Censorship, value theft, surveillance, extraction, extortion etc, the threat list is long. Especially, BPs stand alone as easily identifiable, vulnerable companies that can be pushed around by threats. There are two major groups of threats – inside (above) and outside.
Outside threats are bigger agencies with more money and more resources to craft the outcome that they want. An outside agency can force local BPs to act in its direction, not the direction of the community. Many examples abound. The agency can be big like a state, medium sized like a criminal gang, or small like an extortionist or mugger. In many countries, there are widespread criminal gangs that prey on those who have value. Violence, kidnapping and murder are common.
DPOS provides one weapon against BPs being attacked – voting them out. But the ‘democratic’ solution is slow and uncertain. In practice it hasn’t worked effectively for EOS as yet: As Thomas Cox says, there are no activist stakeholders. Many important stakeholders are not represented, as e.g., businesses have high working assets and low spare staked EOS. Those who are deep insiders are often holding too much information and too little EOS to sway the voting. Information to the wider market is weak, confused and easily disputed, for every story there is an anti-story.
Dispute resolution provides an independent view over difficulties and by means of rulings delivers strong opinion to the community. In this way it works hand in glove with the stake holding voters – dispute resolution strengthens the DPOS feedback process by giving solid information. Properly formed dispute resolution can speedily order remedies and thus make the BPs safer against cronyism and false flag attacks.
Dispute resolution takes away the right to make changes to the blockchain via means not acceptable to the code. This means that a BP can simply say, “we don’t have that right, file a dispute.” Which happened in the Lost Key Cases – BPs quite correctly exercised their rights under the Constitution to refer disputes to the forum of dispute resolution.
Not having that right under contract is a powerful weapon against powerful adversaries internal and external. Coupled with the voters power to remove, the two can work well together.
7/ Reduction in Workload & Cognitive Stress
BPs have to do extraordinary things anyway when the code breaks. Because they are only human, they will be subject to continual pressure to do this and that and another special thing for a special friend. Unfortunately they won’t have time nor patience to deal with the flood of requests. In the end, there will be two outcomes:
- emergency recovery is only available to high value disasters of high value friends of BPs, as happened to Ethereum OR
- BPs will (re)invent dispute resolution as an independent decision making body, so as to cut their workload, and make the system of interventions fair.
BPs fundamentally do not want to decide on issues of rights. They are too hard. To argue whether a lost key case can be resolved is a community issue not a technical block production issue. The design of EOS takes that sort of high level rights questions out of their hands because, in part, it’s simply too hard to rely on BPs to both produce blocks and interpret the rights of the community and keep their own interests out of the reasoning and keep themselves safe from interference.
Are there any circumstances in which this is NOT so?
Above is a long list of benefits from the design of separation of concerns inherent in EOS. But no plan works for ever. What can possibly go wrong?
The detailed thing that can go wrong is a bad ruling. The big picture thing is that the forum of dispute resolution can start acting against the interests of the community. See post on checks and balances for many ways in which the system defends itself against these woes.
Where we are at the current time is #7:
7/ A Further Control – the parties can decline to follow the order. In which case they are in default. Now what happens next is difficult to predict because it depends on circumstances, but let us imagine that the ruling has a defect of seriousness.
In that case, the BPs may decline to enforce the order. This is Most Serious! Because the BPs are the first among equals – they are the ones that enforce the code, the Constitution, and all such good things on the community. If the BPs feel so outraged that they were to stare off the forum, and the forum was so adamant as to stare off the BPs, then we would have a stand-off – and when two of the powers are in disagreement we have constitutional crisis. That would have to be resolved by the community.
EOS then is blessed with that old Chinese (*) curse – to live in interesting times.
(*) while known in the english world as a Chinese curse, there is apparently no known equivalent in the Chinese history.
原文链接/Original URL:
