The science is real. The math is published.
A deterministic, verified system has zero informational entropy. Every state known. Every path verified. That is a statement about information, not about heat. And it is what our compiler proves.
[01] FORMAL DEFINITION
Digital Circuitality
Shannon (1948) defined the entropy of a discrete source as H(X) = -Σ p(x_i) log_2 p(x_i). When a system is completely deterministic — every input produces exactly one output through every path — the probability distribution collapses to a Dirac delta on the correct result: H(X) = 0. This applies to the circuit’s formal specification. Runtime execution introduces implementation-level variance that the specification abstracts away.
A system S exhibits Digital Circuitality if and only if:
Full coherence means zero informational uncertainty
When Φc certifies a system, every state is known and every path is verified
Φc = 1 certifies that every input domain is bounded, every operation verified, every output range proven, and no execution path is undefined. There is no informational uncertainty because there is no unknown state.
Conventional software operates with informational uncertainty > 0. Unverified execution paths, unexplored states, unbounded inputs. Testing reduces informational uncertainty but never eliminates it — Dijkstra (1976): “Testing shows the presence of bugs, never their absence.”
Digital Circuitality eliminates informational uncertainty by construction, not by sampling.
[02] THERMODYNAMIC ANALOGY
The analogy and its limits
The term “thermodynamic” in Digital Circuitality is an analogy, not a physical claim. A physical circuit is coherent when energy flows from source to sink without leaks, all signal paths are closed, and the circuit reaches steady state.
| Physical circuit property | Computational metric |
|---|---|
| Energy flow without leaks | Transfer efficiency |
| Closed signal paths | Circuit closure (Φc) |
| Signal integrity | Signature verification |
| Full connectivity | Verification completeness |
| Circuit complexity | Structural complexity metrics |
What the analogy does NOT claim: No physical energy cost from Φc = 1. No equivalence between computational and thermodynamic coherence. No claim that thermodynamic laws govern compilation.
[03] EVA ALGEBRA
Composition operators
Sequential
Do A, then B. Output of A feeds input of B.
Parallel
Do A and B independently. No data dependency.
Conditional
If X then A, else B. Both branches verified.
Each operator preserves the correctness of its operands. If Part A works and Part B works, their composition is guaranteed to work. This is what hardware has always had — and software never did.
[04] COHERENCE METRICS FRAMEWORK
CMF: Three metrics, one condition
Circuit Closure
Certifies that every branch has a complete input-to-output path. No dangling operations. The computational analog of a closed electrical loop.
Signal Integrity
Verifies that observed behavior matches the expected specification exactly. The analog of signal integrity — no distortion, no noise.
Verification Completeness
All paths have been verified. No execution path has unknown behavior. The analog of full connectivity in a circuit.
Certification Condition
Certification is binary. All three conditions — closure, integrity, and coverage — must hold simultaneously. There is no partial certification.
If any condition fails, the program does not compile.
[04.5] CLOSURE THROUGH DOMAINS
Closure through domains
Circuit closure (Φc = 1) requires that every execution path produces a valid, bounded result. Domain constraints are the mechanism that makes this possible.
1. Input domains bound what enters the circuit. Every input declares a numeric range. Values outside the range are rejected at compilation.
2. Arithmetic on bounded inputs produces bounded outputs. This is provable: if a ∈ [0, 10] and b ∈ [1, 5], then a + b ∈ [1, 15] and a / b ∈ [0, 10].
3. The compiler traces bounds through every operation to verify closure. Each intermediate result inherits a computed domain. If any operation could produce an unbounded or undefined result, the compiler rejects the program.
4. If any path can produce an unbounded result, compilation fails. There is no runtime fallback. The circuit either closes at compile time or it does not exist.
Hardware Analogy
This is analogous to how hardware engineers specify voltage ranges for every pin. A logic gate rated for 0–5V does not accept 12V — the domain IS the specification. In BRIK64, the same principle applies to every variable in every circuit.
[05] INFORMATION THEORY BASIS
Informational entropy ≠ thermal entropy
Recent research in information physics has demonstrated that informational entropy and thermal entropy are fundamentally different quantities. Treating them as interchangeable is a category error.
For Digital Circuitality, the consequence is direct: the coherence framework measures informational entropy, not thermal entropy. No thermodynamic claims are needed for the framework to be rigorous. The verification operates on purely informational foundations.
[06] BRILLOUIN CONNECTION
From Landauer to Brillouin as inspiration
Digital Circuitality draws conceptual inspiration from Brillouin’s work on the relationship between information and entropy, while operating on purely informational foundations grounded in Shannon’s framework.
The system does not depend on any physical thermodynamic claims. It acknowledges the historical inspiration while maintaining rigorous separation between informational and physical domains.
[07] DETERMINISTIC VERIFICATION
Zero informational uncertainty by construction
A formally verified, deterministic system has zero informational uncertainty. Every state is known, every path verified, every domain bounded, the circuit is closed.
This is what Φc= 1 means in Digital Circuitality: the system’s informational entropy is zero — not by testing, but by mathematical construction.
[08] UNIVERSAL TRANSPILATION
Transpilation through informational closure
Traditional transpilers operate at the syntactic level: parse an AST in one language, emit an AST in another. BRIK64 operates at the semanticlevel — extracting the computational essence (what it computes, not how it’s expressed) and encoding it as a PCD circuit.
The critical property: if two programs in different languages produce the same PCD circuit, they are functionally equivalent. PCD captures the informational content of computation independent of syntactic vehicle.
The arithmetic
- • 10 input languages → PCD → 14 output targets
- • 10 + 14 = 24 components for 10 × 14 = 140 transpilation paths
- • Same architectural idea as LLVM (frontends + IR + backends)
- • The addition LLVM doesn’t have: formal equivalence certification
The TCE certifies that the PCD circuit is closed (Φc = 1), guaranteeing the computation is deterministic, total, and informationally preserving. The equivalence is algebraic, not tested.
[09] REFERENCES
Academic foundations
Shannon, C.E. (1948)
A Mathematical Theory of Communication.
Bell Syst. Tech. J. 27, 379–423
Foundation: informational entropy, the framework in which the entire system operates.
Brillouin, L. (1953)
The Negentropy Principle of Information.
J. Appl. Phys. 24, 1152–1163
Inspiration: conceptual information-entropy connection. Not foundation, inspiration.
Dijkstra, E.W. (1976)
A Discipline of Programming.
Prentice-Hall
Motivation: “Testing shows the presence of bugs, never their absence.”
Kish, L.B. et al. (2016–2018)
Research on the distinction between informational and thermal entropy.
Journal of Computational Electronics
Foundational: informational entropy and thermal entropy are distinct quantities.
ACKNOWLEDGMENT
Prof. Laszlo B. Kish(Texas A&M University) reviewed the foundational theoretical framework of Digital Circuitality. His research on the distinction between informational and thermal entropy informed the theoretical foundations of the system.
PUBLICATIONS
Academic publications
arXiv preprint
Digital Circuitality: A Formal Framework for Verified Software Composition
Submitted to arXiv (ID 7419107) — cs.PL + cs.AI + cs.AR
128 monomers, 110,000+ tests, 14 compilation targets, mathematically certified
Zenodo Archive
BRIK64 Papers I-IV v3 + Mathematical Proofs
DOI: 10.5281/zenodo.19313743 — Permanent academic archive
207 mathematical proof files, 0 admits, complete formal verification
The logical chain
1. Shannon (1948) establishes that deterministic systems have zero informational entropy
2. Modern research confirms informational entropy is distinct from thermal entropy
3. A deterministic, verified computer has zero informational uncertainty
4. BRIK64 builds a compiler that certifies this property by mathematical construction
5. The BPU materializes this certification in silicon, where verification is physical and non-maskable