The provable conditions that a contract will conclude securely in the instruction cache. Kernel guarantees calls to identified ledgers, state erased, kernel has hardware secure personal memory. We can add, a validation link back to the foundry.
If we can prove that, then we also prove the contract is valid under trusted miners with a measurable scofflaw leak. We can have this because we can adjust timeouts, which also exist in the spectre complete version. So we always have cancel within timeout on ledgers, which should contain the scofflaws. In other words, a defined protocol on contracts, restriction in their design to the set of finite provables. It is a huge set, and modifiable in many ways via unique ledger systems. But it is rule based finite spanning tree, small N; computable on the spot almost anywhere, a sandbox contract format, hardly a language. An abstraction of the instruction cache under ledger control, the contract as a machine of finite duration.
My chosen incarnation would be conditions on accounts in spreadsheet form. But graph constructs work, set groupings, histograms. All about rules for passing points of agreement in a contract. The interpreter runs under a secure ID. So, to a great advantage, all parties get comfortable with certain, accepted known contracts, the simplest being no double spending, and all parties know all parties obey the rule, to within a counterfeit. The old Nash equilibrium. That is the main point, get it down to that one counterfeit problem, the one exogenous technology. Then it is simply, how secure the foundry.
No comments:
Post a Comment