>How do these smart contract deal with the real world? I can see how they can work for things that entirely involve activities that take place on the block chain (e.g., a smart contract that automatically pays a crowd funded project if and only if it meets a threshold for pledges by a deadline, and refunds the donors otherwise).
They can't deal with the real world by themselves, but where they have value in agreements involving real world events is that they allow the real world assessments to be debundled from those aspects of the agreement that can be formalized.
So in the cotton shipment example, the smart contract would have three parties assigned as Oracles that determine whether the correct ship arrives at the correct port on its correct trip. The smart contract would dictate that if 2 out of 3 of the Oracles agree that the condition has been met, then the smart contract will transfer the funds. There could also be a failsafe clause in the smart contract, that is controlled by a set of five oracles, who have the power to override the smart contract if the majority deem that an expected event occurred.
So yes you need human intervention, but only for those parts that cannot be automated. This explicit definition of what the human roles are may encourage better contracts, by forcing the counterparties to spend more time reasoning about which aspects of the smart contract are subject to ambiguity.
They can't deal with the real world by themselves, but where they have value in agreements involving real world events is that they allow the real world assessments to be debundled from those aspects of the agreement that can be formalized.
So in the cotton shipment example, the smart contract would have three parties assigned as Oracles that determine whether the correct ship arrives at the correct port on its correct trip. The smart contract would dictate that if 2 out of 3 of the Oracles agree that the condition has been met, then the smart contract will transfer the funds. There could also be a failsafe clause in the smart contract, that is controlled by a set of five oracles, who have the power to override the smart contract if the majority deem that an expected event occurred.
So yes you need human intervention, but only for those parts that cannot be automated. This explicit definition of what the human roles are may encourage better contracts, by forcing the counterparties to spend more time reasoning about which aspects of the smart contract are subject to ambiguity.