Mike Hearn: Why Bitcoin Forks

Yes, it is coming. The community is starting to separate, and Bitcoin is about to fork: the software and perhaps the blockchain. The two sides of the split are Bitcoin Core and a minor variant of the same program, called Bitcoin XT. As of August 16, Beijing time, there is now a complete version of Bitcoin XT.

This kind of fork has never happened before. I want to explain this from the perspective of a Bitcoin XT developer: it cannot be said that there was not enough communication.

Bitcoin fork, this topic may make many people curious, so this article is written for ordinary readers. It will not involve the knowledge that has been debated before.

The original version of Bitcoin, as laid out by Satoshi Nakamoto, has always been very clear. The controversy is about growth. In 2008, he answered the first question about Bitcoin's design, saying:

Visa processed 37 billion transactions in fiscal year 2008, or an average of 100 million transactions per day. Such a large number of transactions requires 100GB of bandwidth = 12 DVDs or 2 HD quality movies = bandwidth at the current price of about $18.

Assuming the Bitcoin network reaches this size, it will take years, by which time sending 2 HD movies over the Internet may not be a large transaction anymore.

At the time, he (Satoshi Nakamoto) was more annoyed than any of us about the scaling issues of Bitcoin. His plan was to make Bitcoin popular from the beginning, and he knew that this success would change how people used his system. In 2010, he said:

“We keep the (blockchain) file as small as possible, which is good.

The final solution will not care how big it (the blockchain file) gets.

But for now, while it's still small, keep it small so new users can grow faster. When I eventually implement client-only mode, it won't be a problem anymore.
”

In 2011, I expanded on Satoshi’s scaling intuition in detail through a series of calculations: If Bitcoin becomes so popular, will it completely replace VISA? The answer is that his plan is credible - you don’t need anything else besides a computer, even with such a large amount of traffic. Before he left, I also implemented the model he talked about.

Satoshi's project brought us together. It has changed the lives of thousands of people around the world. Some of us gave up our jobs, others dedicated their spare time to this project, some founded companies, and even traveled around the world. It was an idea of ​​ordinary people being able to pay each other through the blockchain and create this global community.

This is the vision I signed on to. This is the vision Gavin Andresen signed on to. This is the vision millions of developers, startup founders, evangelists, and users around the world signed on to.

And that vision is now in jeopardy. In recent months, it’s become clear that there is a small group of people who have completely different plans for Bitcoin. These people never really understood Satoshi’s intentions because they fear success. What if the technology never improves? What if people can’t run Bitcoin on their home computers? Doesn’t that make Bitcoin less decentralized and more like banking? What if people start to rely on Bitcoin, even if it’s imperfect?

Now that Satoshi Nakamoto has chosen to disappear, they want to make major changes: significantly increase transaction fees, end support for mobile P2P wallets, abandon unconfirmed transactions, and many other things that were not in the project's founding documents.

The so-called “ Lightning network network,” which is being promoted as a replacement for Satoshi’s design, does not exist. The whitepaper describing it was announced earlier this year, and if it ever comes to fruition, it will be a huge departure from the Bitcoin we know and love. To pick just one of the many differences, Bitcoin addresses don’t work. What they will be replaced with has yet to be worked out (because no one knows). There are many other surprising pitfalls, which I’ve covered in another article. What it will ultimately produce that will make our existing Bitcoin network better is still extremely unclear.

What happened to the free market?

In theory, none of this should be a problem. The Lightning Network is built on top of the blockchain, but it requires a fairly trivial upgrade process to achieve its best functionality. Of course, people are willing to explore this direction, and that's totally fine. If what they build works better than the existing regular Bitcoin network, then the market will choose their way, and if that's the case... fair game for them! Bitcoin's current design is unlikely to be the final version for payments. It's reasonable to imagine that one day it will be eliminated in the competition, or enhanced by something else.

But our system, it works today, it has an ecosystem, it has developers, exchanges, wallets, ATMs, books, apps, conferences, and a lot of people have already learned how it works.

Given a free choice, would people decide to migrate to a completely different system?

We don't know, and the people who are pushing this stuff don't want to let the market decide, and that's what's going on.

A long, long time ago, Satoshi set up a temporary "knack computer": he limited the size of each block to 1 MB. He did this to keep the blockchain small in the early days, until the creation of what we now call SPV wallets (which Satoshi called 'client only mode'). As mentioned above, it can be adjusted when the time comes, and it was never meant to be permanent. In the end, it will not matter. I wrote the first SPV tool in 2011 together with my esteemed colleague Andreas Schildbach, and together we built the first and currently most popular Android wallet. Since then, SPV wallets have been used on all major platforms. So Satoshi's reason for this temporary limit was solved a long time ago.

As Bitcoin continues to grow, its blocks are also getting bigger. Reasonable traffic forecasts show that blocks will reach the limits of the current system sometime next year, or at the latest in 2017. If another bubble or stress cycle forces us to exceed this limit before then, the results may not be pretty.

So, now is the time to raise the cap, or remove it altogether. That’s the plan, and that’s where the trouble begins: those who don’t want to see Bitcoin scale have decided to delay it. They see a wonderful, one-time opportunity to forcibly divert Bitcoin from its intended path to an entirely different technological trajectory. They don’t know what that alternative design will be, and they certainly haven’t built it yet. But that doesn’t matter. They think that by blocking the growth of the blockchain, they can “incentivize” (i.e. force) the Bitcoin community to switch to something different, something more to their personal technological tastes.

Why limit blockchain?

So far, I haven't said much about my views on these people, or who they are. I feel that naming names in this article is a very time-consuming and laborious task, and in the end it seems futile. Those who care about this matter must already know, and those who don't know can't recognize the people involved in this matter.

Suffice it to say that they are among the very few who have access to the Bitcoin Core codebase, or those who are convinced by their arguments.

So we won't discuss these debates here, there are too many. For every question that has been raised, Gavin and I have written articles analyzing them and refuting them. Sometimes the answers are common sense, and sometimes they are deeper and require more work, such as network simulation experiments.

The best place to read about these controversies is on Gavin's blog. I was hoping to find a link to a similar collection of opinions refuting Gavin's views, but there isn't one.

To summarize the long and arduous debate, there are several different opposing groups:

1. If Bitcoin approaches this limit, we will be incentivized to create something better.

2. The limit should be raised, but it is not ready yet (the actual time is unspecified).

3. If Bitcoin scales, it will become more centralized, and then it will no longer be Bitcoin.

For others: If you support an objection that is not listed above, please check Gavin's blog and find the answer to it.

The first point may become a reality someday, but it is nothing compared to a theoretical system on paper. But no one who has seen any alternative solutions on the table would think that they will be available in 12 months, (see the last paragraph for another example)... even assuming they are better. This is also an example of the Nirvana Fallacy:

The Nirvana Fallacy is an informal name for the fallacy of comparing something to something that is unrealistic, an idealized alternative. It can also refer to the tendency to believe that there is a perfect solution to a particular problem, and is therefore also known as the perfectionism fallacy.

By creating a false dichotomy, one option is clearly advantageous. But it is also completely implausible. People who use the Nirvana Fallacy can attack any opposing idea as imperfect. According to this fallacy, the choice is not between real-world solutions. One is a realistically achievable solution, and the other is an impractical solution, and it is a "which is better" choice between the two.

The answer to the second objection is that it is too vague. There is reason to believe that a full upgrade of every Bitcoin node could take a year, and that overloading the actual Bitcoin network capacity would cause serious damage. We really should be preparing for this long before then. Two people on the Bitcoin development mailing list who have professional experience in capacity planning believe that this process must begin immediately. Postponing it to some indeterminate future is not a sound engineering practice.

The last point is the most troublesome and still the most controversial. It is based on two assumptions:

1. If Bitcoin can grow, it will grow, and it will grow faster than technological advances that reduce its costs.

2. If the Bitcoin network were tiny, it would be more “decentralized” in the abstract sense than a successful, global Bitcoin network.

If the Bitcoin network was compelling, demand for it would be unlimited: I could stop improving my software and wait for the price of Bitcoin to rise and become rich. Back to Bitcoin as it exists today, it is in a competitive market. Growth is not a God-given right. Every user needs to make an effort, and convincing everyone takes time. Bitcoin is currently growing, but only at a gentle pace. I wish I could confidently say that the cost of running a full node will go up in the future: this would mean that our success is the result of the combined efforts of the entire hardware industry. This industry would be amazing: the iPhone came out in 2007 and cost $500, and only seven years later, the new P9 smartphone appeared, which cost only $30 and matched every feature of the iPhone and significantly exceeded it.

The second assumption goes to the heart of the controversy: Should Bitcoin grow, even if that results in changes to the structure of the network?

In the current system, every user is a network node, but this does not mean that they are the nodes required after the system is large-scale. This is like every newsgroup user running their own NNTP server. This design makes users just users.

- Satoshi Nakamoto, July 2010

To this question, the founder of the Bitcoin project gave a clear answer - YES

We have been working on this plan to this day. Trying to change the answer to this question to NO would not only violate the social contract of Bitcoin, but also go against the wishes of many people in the community.

Those who truly believe an über niche currency would be better should create an altcoin with a limited blockchain size.

indecisive

Why can't this dispute be resolved in a more civilized manner, instead of ending in a complete split?

Simply put, resolution regarding Bitcoin Core’s decision-making process is broken.

In theory, like all open source projects, the core will have "maintainers". The maintainer's job is to guide the project, what should happen and what should not happen. The maintainer is the boss. A good maintainer needs to collect feedback, weigh arguments, and then make decisions. However, the debate over the block size of Bitcoin Core has been delayed for several years.

The problem is that any change, no matter how obvious, can be completely vetoed if it becomes "controversial." With five maintainers, and many other non-maintainers who can also "dispute," a stalemate has been created. The fact that the block size was never permanent is no longer important: removing the block size limit itself is what is being debated. It's like a committee without a chairperson; the meetings never end. To quote one maintainer, "Bitcoin needs a leader like a fish needs a bicycle."

So what about everyone else?

The wider community is absolutely not interested in this issue.

The proposal to increase the block size has support from the following people, among others:

1. Developer of the most popular iOS and Android Bitcoin wallets, and one of the most popular web wallets with millions of users.

2. Several of the largest Bitcoin exchanges.

3. The two largest third-party payment processors occupy the vast majority of the third-party payment market.

4. Several major mining pools, including all Chinese mining pools.

5. Two of the top five Bitcoin Core maintainers (Gavin and Jeff).

6. Polls by users in online forums showed that about 75%-80% supported capacity expansion.

This list is far from complete. Many key players in the ecosystem, who have not yet commented publicly, have expressed support for Gavin and me privately.

So there's nothing weird about what Gavin and I did. If we didn't make this change, someone else would have.

How do those who pushed for 1MB blocks respond to everyone?

They don’t, because they have never been asked to answer their opinions. To quote an influential member of the Bitcoin Core community:

"And by the way, there's a risk in using a company as a megaphone, you can end up with a consensus of misleading opinions."

Companies arguably represent the most passionate, engaged, and technical people in the Bitcoin world. They provide important infrastructure, yet the opinions of those who created them can be considered “misleading consensus.”

What about wallet developers? They are the ones who have the most contact with users’ daily needs.

Don't ask when they will speak up, it's meaningless, their opinions are irrelevant.

This is not surprising: it is becoming increasingly clear that the “consensus” that the Bitcoin Core community often refers to really means the opinions of a few people, regardless of what others outside the community think, what work they have done, or how many users their products have.

In other words, “developer consensus” is marketing that pulls a wool over the eyes of Bitcoin users, making them blind: it only takes two or three people acting in unison to change Bitcoin in any way they see fit.

Are they aware of the opposition from a large number of key people? No, to quote one of the maintainers again:

Gavin is pretty much alone in the Bitcoin Core technical community on this. Many people in my company are concerned about the viability of Bitcoin as a decentralized system, and most of these concerns about this product are being addressed in the technical community.

This is only true if you categorize the vast majority of engineers building the Bitcoin ecosystem as “non-technical.”

How to resolve arguments?

Clearly, the problem has gotten serious. Communication has broken down, and both sides feel they are protecting Bitcoin’s decentralization and are the true vision for Bitcoin. The community has split.

This leaves one last dispute resolution mechanism. We could make a modified version of the software that lets miners vote on whether to upgrade, using the normal chain fork logic. If the majority upgrades to the latest version and produces a block larger than 1MB, the minority will reject it and put it on a parallel chain. To get back in sync with the rest of the network, they will have to accept the fork and support this system. If the majority chooses not to upgrade, then the fork will never happen and the 1MB limit will remain.

It seems that all participants should support this approach: obviously, consensus can no longer be reached through normal mechanisms, so conducting a vote-like action should be a good way to make progress.

Here comes the final and most fatal source of disagreement.

Of the five Bitcoin Core maintainers, Gavin and Jeff support the fork, but the other three believe that any contentious hard fork is unthinkable, crazy, reckless, and should never happen, which would seriously damage Bitcoin and could even be fatal. If any of them were in favor of solving the problem through a fork, I would not have seen them express this.

We strongly disagree with this assessment. We do not think that the sky will fall if the chain forks. We think that people on the smaller blockchains will choose to upgrade and continue to work on the larger blockchain. They will have enough time to learn about the change and prepare for it. Of course, this makes no difference to the Bitcoin Core developers who like or dislike hard forks. They are always firmly against it.

In short, they believe that the only mechanism that limits them should never be used (referring to computing power voting).

I don't think this is an accident, but it is. Their view is that there should be no alternative view to their decisions. Anyone who disagrees with them, for whatever reason, should be blocked forever... Then Bitcoin becomes their toy.

This situation cannot continue. The Bitcoin Core project has shown that it cannot be transformed, so it (the old chain) must be abandoned.

This is why Bitcoin forks.

We (referring to authors Mike Hearn and Gavin et al.) hope that everyone can understand this.

－－－－