secondary title
It all starts with the pursuit of the perfect bottle of Pepsi
In the 1970s, a guy named Howard Moskowitz, who worked in surveying, was a consultant to the food industry. One of his early customers was PepsiCo, which approached him to help determine the dose of aspartame to put in its new Diet Pepsi drink. To get around this, Howard made a batch of Pepsi with different unit doses of the sweetener. He put people from all over the country through numerous tests and recorded their responses.
When he tabulated the data from these tests, he expected to find a cluster of data around a value that obeyed a normal distribution that would reveal the answer to the question—the dose of aspartame. Unexpectedly, the data he retrieved was a mess, and everyone's preferences were not clustered together like a random distribution. These results were not surprising to people at the time, they just felt that data is sometimes not suitable for answering such questions.
secondary title
Prego Road Show
Afterwards, when Campbells hired Howard to improve the recipe for Prego's pasta sauce, he was ready. Howard solved this problem by creating a variety of test sauces that varied in many ways, including spiciness, garlic, levels of visible solids, and more. He came up with 45 different sauces, then tested all of them with people on the streets, all over the United States.
He collected a large amount of data but when analyzing the data, he did not look for that ideal value, but looked for data clusters about people's taste preferences.
He found clear clusters of data around a number of different formulations such as: mild, spicy and tangy. Among them, he found that size was the most important, and that this went against the prevailing thinking about the shape requirements of pasta sauce at the time. Traditional sauces should be thin with no visible solids.
secondary title
Why am I telling such a long list of stories?
To me, the crux of the story is this: When people think of Pepsi and spaghetti sauce, they think there's only one best version. Such an idea should be a so-called Platonic ideal. What the previous companies had to do was to figure out this ideal value and rack their brains on how to make the product as close to the ideal value as possible.
As it turns out, this was completely wrong. There is no ideal version of these things. There is no single best solution. In fact, different versions may be the best choice for different people and different situations.
secondary title
Application on Blockchain
When it comes to blockchain, this is what comes to mind every time I hear messages like: "Our design is the best because we have the most transactions per second", "...because we have the most good governance system”, “…because we are the best decentralized”, etc. It can take more subtle forms such as: What are the right design trade-offs to make when building a blockchain? Where is the sweet spot in the blockchain impossible triangle? The problem assumes the existence of a single optimal or ideal blockchain. When I hear these questions, I like to say, "Who is it best for? For what purpose? From what perspective?
In the process of development and landing, Substrate is very conscientious and cautious about the increasing problems brought about by the core framework and node usage mode. As for Polkadot, it doesn’t require its parachains to be unified, but instead enables them to be tailored to the needs of their specific users and use cases. It tries very hard not to give a so-called black and white universal value.
Dare to admit that there is no most correct method, and there is no unified idealized structure for the ultimate form of the blockchain. This is one of the reasons why Polkadot attracts me, and it is also one of the reasons why I believe in system design. Ultimately, this ability for multiple chains to adapt to individualized specialization, while being able to interoperate to leverage each other's strengths, will be key to Polkadot's long-term competitive advantage.
Answer to the question "What is the best blockchain?": This is originally the wrong question - from Howard Moskowitz's paper, the question should be "What is the best blockchain (they)?"
The answer is that there are likely to be a large number of "best" blockchains out there, as many different user communities and use cases. I expect that over time there will be a multitude of blockchains that cater to different types of needs and specialize for different purposes. I think most of these needs and purposes are things we don't even think about.
But whatever the end result, developers will take a practical approach. They will use frameworks like Substrate to speed up their development efforts, and connect it to networks like Polkadot to expand their reach to other blockchains for users, assets, and specialized services.
compile/sneak
