Arithmetic · Geometry · Computation

Orders, complexes
& quantum
arithmetic.

A mathematical journey from integral lattices to elliptic curves, quantum algorithms, and the analytic questions of Yang–Mills theory.

A didactic monograph, selected Lean 4 proofs, and an expanding atlas of interactive laboratories. Every connection carries its hypotheses; every claim carries its evidence.

New: circuits on a surface →
Execute an 18-qubit toric code, inspect its clean dessin, train a local variational circuit, and examine recorded GPU runs.

VARIETY OF COMPLEXESℝ⁴ → ℝ³
The four-dimensional cone x·y=0 has a torus as its unit link.
x · y = 0A four-dimensional equation,
seen through a three-dimensional projection.
Inspect the construction →
01

Integral foundations

Roggenkamp & Huber-Dyson’s order theory, with source coverage and historical corrections recorded.

02

Explicit constructions

Finite models expose the exact algebra, the geometric choices, and the limits of each connection.

03

Auditable evidence

Prose proofs, numerical checks, selected formal proofs, external theorems, and open problems remain distinguishable.

The fluid dynamics extension

Follow the flow.
Inspect the proof.

The second research pass

Learning, diagrams,
and cosmological inference.

The interactive atlas

Six ways to make
the structure visible.

Change a parameter. Read the invariant. Follow the equation back to the proof.

LABORATORY 01 Arithmetic → geometry

A lattice becomes a torus.

Identify opposite sides of a period parallelogram. The quotient remembers the lattice’s complex structure even when the familiar embedded torus has the same shape.

L = ℤ + ℤτ   ·   E(ℂ) ≅ ℂ/L   ·   z = u + vτ
u-coordinate circlesv-coordinate circlesselected cosetDrag with a mouse or use arrow keys to rotate.
The flat period parallelogram changes with τ.

Current arithmetic

Flat area
Gram matrix
Seam residual

What the embedding means. The displayed map is T(u,v)=((2+0.7 cos 2πv) cos 2πu, (2+0.7 cos 2πv) sin 2πu, 0.7 sin 2πv). It visualizes the quotient topology. It does not preserve the flat metric or complex structure of ℂ/L. The seam residual measures periodicity of this embedding.

Read the mathematical connection

A holomorphic group map ℂ/L → ℂ/M lifts to a map z ↦ αz with αL ⊆ M. Thus End(ℂ/L) is the ring of complex scalars preserving L. It is ℤ or an order in an imaginary quadratic field. The manuscript proves this elementary dichotomy and distinguishes it from the quaternion endomorphism algebra of a supersingular curve in positive characteristic.

Sources: manuscript, “Elliptic curves through their endomorphism lattices”; Sutherland, complex multiplication.

LABORATORY 02 Algebra → four-dimensional geometry

The equation that makes a complex.

A column x followed by a row y composes to zero exactly when x·y=0. Its cone in four real coordinates has a torus as its unit link. Rank changes are visible in the same model.

ℝ —x→ ℝ² —yᵀ→ ℝ   ·   yᵀx = 0
α varies, β fixedβ varies, α fixedselected complexProjection can identify distinct 4D points.
Inspect a stratum

Composition residual
Sampled maximum |x·y|
Ranks (d₀, d₁)
Dimensions (H⁰, H¹, H²)

Exact construction, numerical display. Set u=(cos α,sin α)/√2 and v=(cos β,sin β)/√2, then x=ρ(u+v)/√2 and y=ρ(u−v)/√2. Thus x·y=0 and ‖x‖²+‖y‖²=ρ². All diagnostics use these original coordinates. The subsequent 4D rotations change the viewing basis; they are not asserted to preserve the equation in the original coordinate split. Floating-point ranks use tolerance 10⁻¹⁰.

Why the torus appears, and what rank tells us

The orthogonal coordinate change (x,y) ↦ ((x+y)/√2,(x−y)/√2) converts x·y=0 into ‖u‖²=‖v‖². On the unit sphere, both squared lengths are 1/2, leaving an independent circle in each plane. For the displayed three-term complex, dim H¹=2−rank x−rank y. The alternating sum dim H⁰−dim H¹+dim H² stays zero, including at the cone vertex.

Source: manuscript, varieties of complexes and finite-dimensional Hodge theory. The displayed two-vector case is proved directly; it does not visualize every Buchsbaum–Eisenbud variety.

LABORATORY 03 Arithmetic → a finite group

Nineteen points, one cyclic group.

Over a finite field, an elliptic curve is a finite collection of residue pairs together with the point at infinity. Repeated addition of P visits every point in this example.

E / 𝔽₁₇ : y² = x³ + 2x + 2   ·   P = (5, 1)
Axes are residues modulo 17. The point at infinity 𝒪 appears in the separate box. These marks are not a real cubic.

Group size
19
Order of P
19
Discriminant mod 17
4 ≠ 0

The computational question. Scalar multiplication takes k and computes [k]P. The discrete logarithm asks for k given P and [k]P. This nineteen-element group is an educational example. It makes no security claim about cryptographic parameter sizes.

All points and the addition formula

For distinct affine x-coordinates, λ=(y₂−y₁)/(x₂−x₁). For doubling with y≠0, λ=(3x₁²+2)/(2y₁). Then x₃=λ²−x₁−x₂ and y₃=λ(x₁−x₃)−y₁, all in 𝔽₁₇. Division means multiplication by a modular inverse. Inverse pairs sum to 𝒪, which acts as the identity.

Complete cyclic subgroup generated by P
k[k]PCurve residual mod 17

Source: exact enumeration and group-law calculations included in the companion code; manuscript, elliptic-curve arithmetic.

LABORATORY 04 A hidden subgroup → linear information

Fourier sampling reveals an annihilator.

A uniform state on the kernel of f(a,b)=a+db transforms to a distribution on the characters that vanish on that kernel. The surviving line records d.

H = {(−dt, t) : t ∈ 𝔽₇}   ·   H⊥ = {(u, v) : v = du}
nonzero probability: 1/7zero probabilityEach occupied cell has the same probability.

Total probability
Largest off-support probability
Nontrivial useful samples
6 / 7

What is being simulated. The code evaluates a finite Fourier sum classically. It illustrates the ideal hidden-subgroup step with exact kernel states and a known modulus. It is not a quantum-device run, a complete implementation of Shor’s algorithm, or a resource estimate. A coset shift changes Fourier phases but leaves these probabilities unchanged.

Derive the support and recover the slope

The amplitude at (u,v) is 7⁻³ᐟ² Σₜ exp(2πi(v−du)t/7). The geometric sum is seven when v−du=0 and zero otherwise. Its squared magnitude is therefore 1/7 on H⊥ and zero elsewhere. A sample with u≠0 yields d=v/u modulo seven; the origin is uninformative and occurs with probability 1/7.

Source: manuscript, finite Fourier analysis and quantum discrete logarithms. The identity is a direct character-orthogonality calculation.

LABORATORY 05 Homological information → spectral limits

An index can stay fixed as an eigenvalue vanishes.

This finite complex has Euler index zero for every ε. Its middle Laplacian acquires an arbitrarily small eigenvalue, so the index alone supplies no positive lower spectral bound.

d₀ = (ε, 0)ᵀ   ·   d₁ = (0, 1)   ·   Δ₁ = diag(ε², 1)
λ₁ = ε²λ₂ = 1current εThe chart axes are linear; the control resolves small ε logarithmically.

Least eigenvalue
Smallest positive eigenvalue
dim ker Δ₁
Euler index
0

The distinction at zero. For ε>0, the smallest positive eigenvalue is ε². At ε=0, zero enters the spectrum, while the smallest strictly positive eigenvalue is one. The vanishing coercivity constant and the jump in the kernel are different statements. This example explains why an index theorem does not establish the Yang–Mills mass gap.

Read the finite-dimensional Hodge calculation

The composition d₁d₀ is zero. With standard Euclidean inner products, Δ₁=d₀d₀ᵀ+d₁ᵀd₁. Its quadratic form is ε²a²+b², so the kernel is zero for ε>0 and the first coordinate line for ε=0. In the latter case the cohomology dimensions are (1,1,0), whose alternating sum is still zero. An analytic elliptic complex requires differential operators and symbol exactness in addition to this algebraic chain relation.

Source: manuscript, Hodge theory, elliptic complexes, and the separation of index from coercivity.

LABORATORY 06 Local gauge choice → an invariant observable

A plaquette keeps its phase.

Changing the phase at a vertex changes the two adjacent edge descriptions. Their contributions cancel around the closed loop, leaving its U(1) holonomy unchanged.

Uᵢⱼ′ = exp(iλᵢ) Uᵢⱼ exp(−iλⱼ)   ·   W = U₀₁U₁₂U₂₃U₃₀
original edge phasetransformed edge phaseAngles are shown in radians modulo 2π.

Holonomy residual |W′−W|
Plaquette action 1−Re W
Gauge group
U(1)

Scope of this model. This is one abelian plaquette, not a continuum limit or a nonabelian Yang–Mills construction. It demonstrates gauge cancellation exactly; the residual reports only floating-point arithmetic in the display.

Read the cancellation proof and edge data

Write Uᵢ,ᵢ₊₁=exp(i aᵢ). A gauge transformation changes aᵢ to aᵢ+λᵢ−λᵢ₊₁. Summing over the four cyclic edges cancels every vertex phase, so the product W=exp(i Σaᵢ) is invariant. For nonabelian groups the ordered product is conjugated by the gauge element at the base vertex; its trace is invariant.

Oriented edge phases in radians
EdgeOriginalTransformed

Source: manuscript, lattice gauge variables and the continuum construction problem.

Integrated research and completed steps

One manuscript.
Executable constructions.

The arithmetic, covering, learned-channel and observed-sky studies are embedded here. Their proofs and results are in the same current monograph, with existing chapters and laboratories preserved.

Read the unified monograph Document and evidence

Loading the integrated experiments. The same monograph remains available above.

The written research

A connected argument.
Explicit boundaries.

The expanded monograph develops the mathematics behind these experiments and records where the proposed connections require additional theorems.

UNIFIED CURRENT RESEARCH MONOGRAPH

Orders, complexes, cryptography
& Yang–Mills theory

Integral modules and trace pairings lead to quadratic and quaternion orders. Finite complexes lead to rank loci, Hodge operators, and elliptic symbols. Fourier analysis leads to the quantum discrete-logarithm framework. Gauge invariance leads to a separate set of analytic existence and spectral questions.

The paper gives elementary supporting proofs in full and identifies advanced external theorems. It examines proposed algorithmic improvements through their input models, correctness requirements, and resource costs.

Included in the research record

  • Original LaTeX manuscript and bibliography
  • Page-level source review and correction notes
  • Selected Lean 4 declarations and build evidence
  • Exact finite examples and reproducible code
  • These interactive mathematical laboratories
  • Preserved access to the original project
Explore inherited project material →View the GRH research repository →Browse the research extension and code →

Evidence & current status

Know what has
actually been established.

PROOF SCOPE

Source-scoped formal evidence

The original seven-module development retains its dated record of 121 selected declarations. This unified integration separately rechecks 26 declarations, including four new arithmetic/operator contracts. These source-scoped counts are not a claim that every theorem in the full monograph or its external continuum sources has been formalized.

The unified integration record identifies the new source, tests and document. The current source-hashed Lean record contains exact types, axiom reports and per-module logs. The written-claim register maps selected formal components to manuscript statements. The release summary records PDF and implementation evidence. Numerical residuals are separate from Lean proof certificates; the independent upstream Navier–Stokes build remains incomplete.

SOURCE COVERAGE

Historical sources with corrections

All 701 supplied book-page OCR texts were read across the review team, with selected visual checks and explicit coverage records. This does not certify every source proof. The review ledger identifies unread linked material and other limits. The second volume’s corrections inform the interpretation of the first.

Roggenkamp and Verena Huber-Dyson are credited for Volume I; Roggenkamp authored Volume II. The original research corpus remains separately identified.

Yang–Mills · status assessed 10 November 2026

Existence and the mass gap
are distinct milestones.

The Yang–Mills working group reports a Wightman quantum field theory from the four-dimensional Wilson lattice scaling limit. Its existence result was independently assessed on 10 November 2026. Uniqueness of the infinite-volume limit and the mass gap remain open.

This companion explains the reported result in its stated scope. It does not claim a solution of the Clay mass-gap problem or derive one from an index theorem.

Working-group source at the assessed commit →
OpenAI research context, 26 October 2026 →

Primary mathematical sources

Continue into the literature.