@Plasma lays stress on the Fund Recovery as opposed to the Continuous Execution.
Plasma is designed based on preliminary assumptions that off-chain execution will eventually be unsuccessful. Operators are able to stop, block transactions, or deny data. Plasma does not aim at ensuring constant availability; instead, it considers what happens when the failures do.
In order to take off-chain the execution to reduce the load on Ethereum, the ownership is never transferred. The money is enforced on Ethereum by pre-established exit schemes. In case child chain becomes unreliable, users do not need to wait until recovery or coordination is achieved. They are able to start exits, file proofs and reclaim assets directly on base layer.
This design causes fund recovery to be the main assurance, rather than continuous implementation. Failure can have an intervening effect on normal activity, but the power to withdraw is maintained. Such a tradeoff determines the structure of Plasma and constrains its expressiveness, as well as outlines its place in the scaling space. @Plasma only considers scalability acceptable in cases where the users have a clear and verifiable way to go back to Ethereum.
