Rohonc Engine

A manuscript undeciphered for two centuries compiles directly to categorical assembly code.

The Rohonc Codex (Hungarian National Library, Oct. Hung. 73) has resisted every conventional analysis since its discovery in the early nineteenth century. It contains no known language. Its ~200 distinct symbols have been linked to Sumerian, early Hungarian, Dacian, and a dozen invented alphabets — none convincingly. This repository applies the same methodology that compiled the Voynich Manuscript to its companion case.

The claim is structural, not interpretive: the twelve visual families of the Rohonc symbol set are the twelve categorical primitives of the Universal Imscriptive Grammar. Compiled to IMASM, the corpus runs at zero thermodynamic entropy delta on the same Tri-Phase Flux Register architecture.

---

Visualizations

Full-corpus animated call-graph

Nodes — one node per page section extracted by the Rohonc engine across all 33 pages of the Rohonc Codex. Node size scales with degree. Node color encodes the manuscript's four structural sections as identified by the engine: liturgical (amber), pictographic (green), astronomical (blue), and mixed/undetermined (grey).

Edges — directed edges encoding structural dependencies between page sections: shared visual-glyph families, sequential co-occurrence patterns, and grammar rule reuse as identified by the Rohonc engine's 12-opcode IMASM compilation. An edge u → v means section u's compiled grammar rule set is a structural prerequisite for v's.

Cross-page back-edges — edges that travel backward in page order (from a later page to an earlier one), forming cycles in the manuscript graph. These encode recursive or self-referential structures in the codex — places where the manuscript's visual grammar refers back to an earlier section. Back-edges flash purple on first appearance.

Phase 1 — build: Pages appear in manuscript order (page 1 through page 33). Forward dependency edges are drawn as each page's node is added. Back-edges flash purple when first revealed, marking cross-page recursion. The title bar shows the current page and its identified section type.

Phase 2 — flow wave: A Gaussian pulse (σ ≈ N/6) travels through the page sequence, wrapping cyclically. Near the pulse peak, nodes enlarge and brighten toward white; edges connecting active nodes glow with increased linewidth and alpha. The title bar shows the current pulse position and the Frobenius identity μ∘δ = id.

---

The twelve primitives

RTFF code	Glyph family	Opcode	Mnemonic	Operation
cr	cross/crucifix	0x0	VINIT	Initial object ∅
hk	hook/fishhook	0x1	TANCH	Terminal anchor ⊤
fa	forward arc	0x2	AFWD	Morphism →
ba	backward arc	0x3	AREV	Contravariant inversion ←
lg	ligature	0x4	CLINK	Composition ∘
lp	loop/mirror	0x5	ISCRIB	Identity id
br	branch/fork	0x6	FSPLIT	Frobenius co-multiplication δ
cv	convergent	0x7	FFUSE	Frobenius multiplication μ
vt	vertical stroke	0x8	EVALT	Lattice: True
hz	horizontal stroke	0x9	EVALF	Lattice: False
cl	closed loop	0xA	ENGAGR	Lattice: Both (paradox)
dt	dot/point	0xB	IFIX	Linear tape write

The mapping is structural: the cross family (initiating, framing) is VINIT; the dot (terminal, irreversible) is IFIX. The branch/convergent pair is FSPLIT/FFUSE — the Frobenius co-multiplication/multiplication that governs the manuscript's pictographic fork-and-rejoin topology.

---

Four sections, four topological regimes

Section	Pages	Dominant primitives	Topology
Liturgical	p1–p50	cr lp vt	Identity-loop chains (prayer repetition)
Pictographic	p51–p150	br cv	Fork-fuse networks (narrative branching)
Astronomical	p151–p300	cl fa	Closed-loop cosmological circuits
Mixed	p301–p448	all	Full-spectrum register saturation

---

Quick start

git clone https://github.com/umpolungfish/rohonc-engine
cd rohonc-engine
uv sync
uv run python examples/quickstart.py

---

Command-line

rohonc-compile  data/rohonc_rtff_sample.txt --log full_log.txt
rohonc-run      data/rohonc_rtff_sample.txt --steps 10000
rohonc-graph    data/rohonc_rtff_sample.txt --output rohonc_graph.png
rohonc-sections data/rohonc_rtff_sample.txt --output-dir rohonc_graphs

---

Transcription format (RTFF)

The Rohonc Transcription Folio Format mirrors the Voynich IVTFF:

<p12>
;H> cr fa lp hk vt cl br dt fa lg
;H> br cv fa lp cr dt hk vt cl fa

<p{N}> opens page N. ;H> lines contain space-separated symbol family codes. The script's right-to-left direction is absorbed into register monotonicity — every instruction advances the PC forward through the compiled stream regardless of source direction.

---

Repository structure

rohonc_engine/       Python package
  primitives.py      — twelve symbol family → IMASM opcode definitions
  compiler.py        — parallel page compiler (RTFF format)
  runtime.py         — Tri-Phase virtual machine
  callgraph.py       — register flow graph generator
  sectional.py       — per-section graph renderer
data/
  rohonc_rtff_sample.txt  — sample RTFF transcription
docs/
  ROHONC_MANUSCRIPT.md    — full technical paper
examples/
  quickstart.py           — full pipeline demonstration

---

Relation to the Voynich Engine

The Rohonc Engine is the second instantiation of the same architecture. The Voynich mapped twelve EVA glyph families to the twelve IMASM opcodes; the Rohonc maps twelve visual-structural families using an identical pipeline. The Tri-Phase VM, register semantics, entropy theorem, and call-graph construction are unchanged. The manuscripts differ in section topology (Voynich: botanical/biological/balneological/cosmological; Rohonc: liturgical/pictographic/astronomical/mixed) and in the primitive surface codes, not in their categorical deep structure.

Both compile to SELF_SUSTAINING_BOOTSTRAP_COMPLETE at entropy delta = 0.00000000 J/K.

---

License

Unlicense — public domain. No conditions, no attribution required.