3 ms·
Yes. They should have gone with something simple and declarative, like decision tables.[1] Those have a finite number of cases and can be exhaustively tested. W
by Animats 6y ago
Yes. They should have gone with something simple and declarative, like decision tables.[1] Those have a finite number of cases and can be exhaustively tested. Which is what you want for a smart contract for something real.
But no, they had to make it Turing-complete. That failed quickly. Remember the DAO debacle. That should have been a teaching moment. But no. Because the people burned were insiders, the whole Etherium blockchain was split to rescue them.
[1] https://en.wikipedia.org/wiki/Decision_table https://en.wikipedia.org/wiki/Decision_table
- intotheabyss 6y agoHow has it failed? Seems to me that turing completeness has allowed Ethereum to flourish.
- FridgeSeal 6y agoYeah which is why I buy coffee in the morning with Ethereum right?
- Sargos 6y agoThinking that buying coffee is an important use case for Ethereum is pretty ignorant. It's literally holding billions of dollars in value and building an ecosystem of companies that work together with a shared purpose. Just because you can't buy a coffee with Ethereum for some reason doesn't mean it's somehow failed.
- pacoWebConsult 6y agoAt the very least people should be formally verifying the behavior of their smart contracts before investing millions of dollars in them. The ETH VM is actually pretty simple to model and formally verify at least some of the basic properties of contracts.
- Animats 6y agoFormally verify against what? You need some simple model of what's supposed to happen to verify against.