Lux(λ) |光尘|空灵|GEB
Lux(λ) |光尘|空灵|GEB|Jul 15, 2025 09:06
The subject of judgment represents the issue of trust. Any programmed component system must have at least one issue that cannot be determined from within the system itself. When designing secure programming components, only one issue that cannot be determined within the system will be left for third-party judgment. Taking Bitcoin programming as an example, any logic in the TX business layer is deterministic, and the self referential determination problem in the TX business layer is handed over to the Block system layer through "computational equivalence" to determine, leaving only one undecidable problem in the Block system layer, which is the existence of the Block itself. The theoretical logic for solving this judgment problem comes from Turing's "Ordinal Logic System", which does not rely on trusting a real deterministic subject (code or person), but instead uses the algorithm of oracle Turing machine+super poor iteration to probabilistically integrate with the time of the real world. Through the time iteration of the physical world, all participants in the Bitcoin network work together to determine the existence of TX/Block. The consensus system coin based on a code system controlled by a finite set consensus person hands over its existence proof to an organization of finite individuals for judgment. And the discriminative nature of tokens programmed based on smart contracts may not be as trustworthy as strict audit rules for stocks if they do not rebuild their own code consensus and instead throw it out for judgment. Therefore, when designing a cryptocurrency system, the first step is to perform a computationally equivalent conversion between the business layer and the system layer, leaving only one undecidable problem within the entire business system, which is the self referential judgment problem. If we want to achieve complete decentralization of the self referential problem, we can rely on the theory of Turing ordinal logic and refer to the Bitcoin method designed by Satoshi Nakamoto. If we want to decentralize the self referential problem based on the trust code (Code is Law), we can delegate the self referential problem to a finite set of consensus agents for control. If the self referential problem is to be entrusted to a trusted person or organization, the best implementation model is the current stock system. The only advantage of Token over the stock system is transparency. So the Turing complete smart contract programming system brings transparency to the encoding, but does not bring security. Security still requires the encoder to redesign the system based on the above principles: that is, to clearly determine all problems in the business layer, leaving no unpredictable problems for the interface, and leaving the judgment problem of the business layer itself to trusted centralized institutions or code runners of finite set agents, or to time ordinal based super poor iterations.
+6
Mentioned
Share To

Timeline

HotFlash

APP

X

Telegram

Facebook

Reddit

CopyLink

Hot Reads