# MrlyProd > https://mrly.net MrlyMath is the mathematics of designs on the corners of a cube: a code says which of the eight corners are filled, the Kronecker product grows that rule into itself level by level, and the object it converges to is a fractal, the Sierpinski carpet and the Menger sponge among them. Everything else on this site is measurement: fills, voids and exposed faces, plane cuts, graph spectra, and the integer sequences the counts write down, with Rust crates doing the arithmetic and the browser only painting. Every claim is proved, checked over a stated finite domain, or labelled a conjecture, and all of it is free and open source. Links are page sources in markdown: /raw/research/P.md is /research/P/, /raw/site/blog/x/index.md is /blog/x/, /raw/site/pages/x.md is /x/; a link outside /raw/ is the page. ## Wiki, in prerequisite order - [The Basel problem](https://mrly.net/raw/research/wiki/basel-problem.md): Add up the reciprocals of the squares and the total stops at pi squared over six, a circle appearing in a question that never mentioned one. - [The difference coarray](https://mrly.net/raw/research/wiki/difference-coarray.md): Put a few sensors on a line and list every distance between two of them. Those distances are what the array can measure, and a well-placed handful of sensors covers far more distances than it has sensors. - [The Euler-Mascheroni constant](https://mrly.net/raw/research/wiki/euler-mascheroni-constant.md): Add 1 and a half and a third and on to one over n, take away the logarithm of n, and the difference settles on 0.5772157, a number nobody has yet placed. - [Euler's number](https://mrly.net/raw/research/wiki/eulers-number.md): Split a year's interest into more and more payments and the yearly growth climbs, slows and stops at 2.718281828. - [Famous formulas](https://mrly.net/raw/research/wiki/famous-formulas.md): Eight elementary rules run out to infinity. Five close on a constant, one counts the primes, one refuses to reach zero and one refuses to settle down at all. - [The Farey sequence](https://mrly.net/raw/research/wiki/farey-sequence.md): Every fraction between zero and one whose bottom number is at most Q, written in order; each new Q slips a few new fractions between the old ones and never moves one. - [The greatest common divisor](https://mrly.net/raw/research/wiki/greatest-common-divisor.md): The largest number dividing two others at once; Euclid finds it by taking remainders, and when it comes out 1 the two numbers share nothing. - [The Kronecker product](https://mrly.net/raw/research/wiki/kronecker-product.md): Stamp a small picture into every filled cell of itself and it grows a level; that one move builds every design on this site. - [The Leibniz series](https://mrly.net/raw/research/wiki/leibniz-series.md): Add one, take away a third, add a fifth, and the running total swings over and under a quarter of pi, closing on it very slowly. - [Parity](https://mrly.net/raw/research/wiki/parity.md): Every whole number is even or odd, and that single bit, read on each coordinate at once, is the whole of what a design uses to choose its cells. - [Prime numbers](https://mrly.net/raw/research/wiki/prime-numbers.md): A prime is a number whose stones make only one rectangle; every other number is built from primes, and the supply never runs out. - [Rajchman measures](https://mrly.net/raw/research/wiki/rajchman-measure.md): Spread a unit of mass round a circle and listen to it at higher and higher frequencies; if the sound dies away the measure is Rajchman, and the Cantor dust is the famous one that never goes quiet. - [Visible lattice points](https://mrly.net/raw/research/wiki/visible-lattice-points.md): Stand at a corner of the grid and some points hide behind nearer ones; the ones you can see are the pairs sharing no divisor, and counting them hands back pi. - [The Wallis product](https://mrly.net/raw/research/wiki/wallis-product.md): Multiply 4/3 by 16/15 by 36/35 and keep going, and the running product climbs forever without ever reaching half of pi. - [Burnside's lemma](https://mrly.net/raw/research/wiki/burnsides-lemma.md): Counting arrangements that turning and flipping should not tell apart, by averaging how many each symmetry leaves untouched. - [Cellular automata](https://mrly.net/raw/research/wiki/cellular-automaton.md): A row of cells, a rule that reads each cell with its two neighbours, and a new row written underneath; run it and the rows draw a picture. - [The digit sum](https://mrly.net/raw/research/wiki/digit-sum.md): The sum of the digits of a number in base b has the same remainder as the number modulo b - 1; apart from that it spreads evenly over residue classes, among all numbers and among the primes. - [The divisor function](https://mrly.net/raw/research/wiki/divisor-function.md): d(n) counts the ways to split n into a product, 2 at every prime and 12 at 60; it never grows as fast as any power of n, and on average it is log n. - [Euler's totient](https://mrly.net/raw/research/wiki/eulers-totient.md): Phi of n counts how many of the numbers 1 to n share nothing with n; it sinks for numbers with many small factors and hits its ceiling at every prime. - [Ford circles](https://mrly.net/raw/research/wiki/ford-circles.md): Above every fraction sits a circle resting on the number line, the simpler the fraction the bigger the circle, and no two of them ever overlap. - [The Gaussian integers](https://mrly.net/raw/research/wiki/gaussian-integers.md): Numbers of the form a plus b times i, drawn as a square lattice, where each point has a size and some ordinary primes break in two while others stay whole. - [Graphs](https://mrly.net/raw/research/wiki/graphs.md): Dots joined by lines and nothing else; count the lines at a dot for its degree, write the joins in a table of ones, and a design's filled cells become a network. - [The Mobius function](https://mrly.net/raw/research/wiki/mobius-function.md): Mu of n is 0 when a square divides n, and otherwise plus or minus one depending on whether n has an even or an odd number of prime factors. - [Moire](https://mrly.net/raw/research/wiki/moire.md): Lay two rulings over each other and they beat; the points both of them mark are the ones their scales agree on, and stacking many scales lights the points most scales share. - [Pascal's triangle](https://mrly.net/raw/research/wiki/pascals-triangle.md): A triangle of numbers where every entry is the sum of the two directly above it, and colouring the odd ones draws a fractal. - [The prime counting function](https://mrly.net/raw/research/wiki/prime-counting-function.md): How many primes are there below a number? The answer is a staircase with one riser per prime, and two smooth curves that chase it without ever quite catching it. - [The Riemann zeta function](https://mrly.net/raw/research/wiki/riemann-zeta-function.md): Add one over every whole number raised to the power s; the answer is zeta of s, it equals pi squared over six at s equal to 2, and it can be rewritten as a product over the primes. - [The Sierpinski carpet](https://mrly.net/raw/research/wiki/sierpinski-carpet.md): Cut a square into nine, throw the middle one away, then do the same to the eight that are left, and keep going for ever. - [The spirograph](https://mrly.net/raw/research/wiki/spirograph.md): A small wheel rolls inside a big ring carrying a pen, and the fraction made by the two radii decides how many petals the pen draws and when it comes home. - [Stern's diatomic sequence](https://mrly.net/raw/research/wiki/stern-diatomic-sequence.md): Write 1 1, then slip the sum of every two neighbours between them, again and again; the numbers that pile up count sums of powers of two, each power at most twice, and, read in pairs, list every positive fraction exactly once. - [The Thue-Morse sequence](https://mrly.net/raw/research/wiki/thue-morse-sequence.md): A string of noughts and ones built by writing a block and then its opposite for ever, which is also the parity of the ones in each place number written in binary. - [Truchet tiles](https://mrly.net/raw/research/wiki/truchet-tiles.md): A square with two quarter circles on it, laid on a grid in one of two turns, joins up into closed loops and long strands that never branch and never cross. - [The Ulam spiral](https://mrly.net/raw/research/wiki/ulam-spiral.md): Wind the whole numbers outwards on a square spiral, light up the primes, and they fall on diagonal streaks instead of scattering evenly. - [The Wallis sieve](https://mrly.net/raw/research/wiki/wallis-sieve.md): Cut a square into nine and drop the middle, cut each survivor into twenty-five and drop the middle, and keep going; the area left is a quarter of pi. - [Beurling generalised primes](https://mrly.net/raw/research/wiki/beurling-generalized-primes.md): Pick any climbing list of real numbers and call them primes; their products are the integers. Beurling's theorem says how straight the integer count must run for the prime number theorem to hold, and the exponent is sharp. - [The Eisenstein integers](https://mrly.net/raw/research/wiki/eisenstein-integers.md): Numbers built from a cube root of one, drawn as a hexagonal lattice, with six units and a size rule that turns the honeycomb into a number system. - [The Euler characteristic](https://mrly.net/raw/research/wiki/euler-characteristic.md): Count the corners, subtract the edges, add the faces; the answer is 2 for anything shaped like a ball, and bending the shape never changes it. - [Goldbach's conjecture](https://mrly.net/raw/research/wiki/goldbach-conjecture.md): Every even number past two is a sum of two primes. Machines have checked it further than anyone can list and nobody has proved it. - [The graph Laplacian](https://mrly.net/raw/research/wiki/graph-laplacian.md): Degree minus adjacency, one table per network; its eigenvalues are the tones the network can ring at, and the staircase they make says how it is knit together. - [The horocycle flow](https://mrly.net/raw/research/wiki/horocycle-flow.md): Fold a horizontal line of the hyperbolic plane onto the modular surface and it closes into a loop of length 1/y; as the line sinks the loop spreads evenly over the whole surface, at a pace set by the zeros of zeta. - [Kummer's theorem](https://mrly.net/raw/research/wiki/kummers-theorem.md): The power of a prime in a binomial coefficient is the number of carries when the two parts are added in that prime's base, so a question about divisibility becomes a question about digits. - [The Menger sponge](https://mrly.net/raw/research/wiki/menger-sponge.md): The carpet's solid cousin: cut a cube into 27, drill out the middle and the middle of each face, then do the same to the 20 cubes that are left. - [The Mertens function](https://mrly.net/raw/research/wiki/mertens-function.md): Mark every whole number plus one, minus one or nothing, then keep a running total. How far that total is allowed to wander is the Riemann hypothesis. - [The transfer matrix](https://mrly.net/raw/research/wiki/transfer-matrix.md): A table that says which state may follow which; multiply it by itself and it counts the walks of any length, and those counts grow like a fixed power. - [Domino tilings](https://mrly.net/raw/research/wiki/domino-tilings.md): Cover a region of the square grid with dominoes, each on two cells that share a side; the number of ways is one determinant, and a random way freezes near the corners and melts in the middle. - [Fractal dimension](https://mrly.net/raw/research/wiki/fractal-dimension.md): Cover a shape with boxes, shrink the boxes, and watch how fast the count grows; the growth rate is the dimension, and for a carpet it is not a whole number. - [The Mahler measure](https://mrly.net/raw/research/wiki/mahler-measure.md): Average the logarithm of a polynomial's size once round the unit circle and you get one number; it sees only the leading coefficient and the roots outside the circle, and for whole-number polynomials it is zero exactly when the leading coefficient is plus or minus 1 and every nonzero root is a root of one. - [Nodal domains](https://mrly.net/raw/research/wiki/nodal-domains.md): A drum skin at one pure tone splits into patches that swing up while their neighbours swing down; the still lines between them are the nodal set, the patches are the nodal domains, and Courant's theorem says the k-th tone has at most k of them. - [The spectral radius](https://mrly.net/raw/research/wiki/spectral-radius.md): Multiply an arrow by the same matrix again and again and it settles on one direction; the factor it stretches by there is the spectral radius. - [The Cantor set](https://mrly.net/raw/research/wiki/cantor-set.md): Cut the middle third out of a line, then the middle third of every piece left, for ever. What survives has no length, as many points as the line, and added to itself fills an interval. - [The random walk](https://mrly.net/raw/research/wiki/random-walk.md): A walker on a graph steps to a neighbour picked at random; after n steps on a grid it is about the square root of n away, and on a fractal it is slower. - [Space-filling curves](https://mrly.net/raw/research/wiki/space-filling-curve.md): A line drawn without lifting the pen can pass through every point of a square, and points that are close along the line stay close in the plane. - [Subshifts of finite type](https://mrly.net/raw/research/wiki/subshift-of-finite-type.md): Write an endless string of letters under a rule that forbids a few short words; the strings that obey form a subshift of finite type, and the number of words it allows grows at a rate read off one small table. - [Substitution tilings](https://mrly.net/raw/research/wiki/substitution-tiling.md): Cut every tile into smaller copies of the tiles you started with, scale the result back up, and repeat; a handful of shapes then covers the plane with the same structure at every scale. - [Minkowski content](https://mrly.net/raw/research/wiki/minkowski-content.md): Fatten a set by a small radius and weigh what it swallows. How fast that weight shrinks is the set's dimension; whether the rescaled weight settles down is decided by the logarithms of its ratios for a self-similar set on the line, while off the line the lattice half is known only under extra conditions on the holes. - [Missing-digit numbers](https://mrly.net/raw/research/wiki/missing-digit-numbers.md): Strike out every whole number that uses the digit 7. Almost all of them go, yet the survivors are a Cantor set in disguise and still hold infinitely many primes. - [Rep-tiles](https://mrly.net/raw/research/wiki/rep-tiles.md): A shape that cuts into a few smaller copies of itself, all of one size, can be cut again and again, and run backwards the same cut grows it to cover the whole plane. - [Wang tiles](https://mrly.net/raw/research/wiki/wang-tiles.md): Square tiles with a colour on every edge, laid so that touching edges agree; whether a given set can cover the whole plane is a question no computer program can always answer. ## Research notes - [MrlyMath](https://mrly.net/raw/research/README.md): The mathematics of the mrly tree: a parity rule on the corners of a cube, substituted into itself by the Kronecker product, and everything that falls out of those two moves. - [Mobius and primes with no two equal adjacent digits](https://mrly.net/raw/research/notes/adjacent.md): The integers with no two equal adjacent base-b digits keep exactly (b-1)^L of the integers with L digits, the dimension of one missing digit, and form no digit design; a run expansion of their transform carries a chain certificate below 1/5 from base 1153, and from there the Mobius function cancels on the set and its primes are counted, A(x)/log x of them, with constant 1, where a design of alternating halves inside the set gave primes only at even bases from 61270 and at a lower count, by a theorem on digit automata of which the Kummer set is the two-state case. - [The Apollonian gasket](https://mrly.net/raw/research/notes/apollonian.md): The circle face of the stack: the Descartes reflection as exact integer arithmetic on (k, k x, k y), the strip packing's line-tangent circles proved to be exactly the Ford circles so the Farey stack is the packing's shadow with brightness floor(Q sqrt(2/k)), the curvature census, root quadruple excluded, whose exponent by ratio lands at 1.305 against the rigorous residual dimension, the eight residues mod 24, and why no design of base at most 100 carries that dimension. - [Loops in arcs](https://mrly.net/raw/research/notes/arcs.md): A design drawn in quarter-circle arcs, its deleted cells turned once, and the closed loops the arcs make level by level. - [Sparse arrays](https://mrly.net/raw/research/notes/arrays.md): A fractal sparse array is a digit design in the balanced base of its generator's coarray, so its coarray weight is a product over digits and, from order 2 on, its fragility is exact: a sensor is essential exactly when every digit has a unique-difference partner, the fragility is (u/L)^r, the power bound of Yang and coauthors is tight exactly when the generator's essential sensors all have such partners, and the bound of Cohen and Eldar exactly when every sensor has one; at a compressed base the essential count can leave the product law. - [The automata](https://mrly.net/raw/research/notes/automata.md): An elementary cellular automaton is a three-dimensional design read as a rule and Life a nine-dimensional one: what the identity buys, and what it cannot. - [What base 3 hides](https://mrly.net/raw/research/notes/bases.md): What base 3 hides: an Eisenstein L-value where base 2 hid pi, and the collapse theorem that turns any multiplicatively dependent family of bases into one design in base r^lcm(e_i). - [Beneath a design](https://mrly.net/raw/research/notes/beneath.md): The object beneath a design as three slots, accept, place and glue: the memory dial priced by a transfer matrix with kappa zero on exactly the non-empty product rules over 19563 live classes, the Collatz carry landed on that dial as a four-state transducer of unbounded radius whose zero-carry sets are the golden and supergolden rules at 3n + 1 and 7n + 1, and the radix dial that keeps every count, carries the gasket, the dragons and the Koch curve as codes with a twist vector, and loses the fill law the moment a digit turns; the Mobius meter of every width-three rule to 2^30 against its own mass, the memory zeta as a matrix ladder with one pole comb per eigenvalue, and the design census up to similarity and up to affine conjugacy. - [A design is a Boolean function](https://mrly.net/raw/research/notes/bijection.md): Designs are Boolean functions up to cube symmetry; the strongest theorem in the tree. - [Two bases](https://mrly.net/raw/research/notes/cobham.md): What a second multiplicatively independent base does to a design: Cobham and Cobham-Semenov at their sources, every proper dim 1 design proved base-locked by density alone, the semilinear designs of dim at least 2 pinned between a proved count condition and the block designs with the diagonal as the standing counterexample, the global planar budget Refuted twice and replaced by one budget per axis for product designs, the base-2 gasket against the base-3 gasket counted exactly to 3^24, and a three-base object on the line whose budget is negative, with five members and no sixth below a height of 38170 decimal digits, and the sumset of the base-3 and base-4 designs {0, 1}, Erdos problem 125, counted to 3^22, its two-base energy ratio proved subpolynomial and unbounded, and its open upper density shown positive once one orbit of an irrational rotation misses, infinitely often, a target of measure at most 190/M. - [Complexity](https://mrly.net/raw/research/notes/complexity.md): Boolean complexity of the catalog, and the Laplacian spectra of the fractals it builds. - [Structure against noise](https://mrly.net/raw/research/notes/connectivity.md): Self-similar designs raced against matched random cell sets on components and boundary. - [The coprimality spine](https://mrly.net/raw/research/notes/coprime.md): The coprimality spine: exact base-local factors on every design, the census behind them, the window at dimension one with the band automaton's critical out-degree and its algebraic block rate, and primes on a design, where the least base below the quarter threshold is 21 at one missing digit and 32 at two, the one-digit family closes at 34 and the two-digit family closes at 21 against the third. - [The core](https://mrly.net/raw/research/notes/core.md): What a design is, the headline counts, and the three genera. - [Crop](https://mrly.net/raw/research/notes/crop.md): The circle on a design: the corner disc count, its log-periodic main term, and the crossing shell read as a rooted tree whose transfer operator is derived from the geometry. - [Cuts](https://mrly.net/raw/research/notes/cuts.md): The six-gasket theorem: a diagonal cut through one parity solid is Sierpinski all the way down. - [Dilated stacks](https://mrly.net/raw/research/notes/dilations.md): The effective receptive field of a dilated convolution stack is a digit-design representation count: Stern's diatomic sequence for three taps at base 2, with Fibonacci peaks and 2L + 1 single-path lags, and a depth limit that is absolutely continuous exactly when the block polynomial vanishes at the right roots of unity, which the two common residual blocks break at every gain but at most one. - [Complex dimensions](https://mrly.net/raw/research/notes/dimensions.md): Complex dimensions: every design in the lattice class, the carpet proved not Minkowski measurable with its explicit profile, and the arithmetic pole at s_0 + 2 pi i/log 3 certified genuine. - [Dimers on a design](https://mrly.net/raw/research/notes/dimers.md): Domino tilings of a plane design: the colour imbalance read off the code, tilings that never switch off, codes that tile late, and the carpet's growth constant bracketed. - [Erdos problems](https://mrly.net/raw/research/notes/erdos.md): The open Erdos problems that are questions about digit designs, ranked by the odds of a checked result and by who reads them, and problem 124 attacked: sums of powers of several bases, where a bitset certificate makes each complete level a finite computation, every base set up to 10 with sum 1/(d-1) > 1 and gcd = 1 is certified at the first four levels, and the two-part route of the literature is proved unable to reach any base set below sum 1/(d-1) = 2, while two window criteria fall to explicit incomplete multisets. - [The stack is an RH-observable](https://mrly.net/raw/research/notes/farey.md): The stack's moire is a Farey resonance diagram; each scale n adds exactly phi(n) bright nodes, the stack is an address rather than a construction (any depth evaluates in closed form, which buys rendering and provably nothing toward RH), and restricting the Farey sequence to a digit design empties a fixed sixth of the line at base 3 {0,1} when both coordinates are restricted, while restricting the denominator alone leaves the Franel-Landau shape reading the same. - [The restricted Franel identity](https://mrly.net/raw/research/notes/franel.md): Franel's identity on the Farey fractions whose denominators lie in a digit design: the exponential sums are dilated Mertens sums, the square-root threshold is equivalent to their square-root cancellation in mean square over the dilates, the dilates are regular languages read by long multiplication, their masses need no Mobius input, and the design's repunits refute the uniform law that would give a rate in d. - [The digit-sum twist on a design](https://mrly.net/raw/research/notes/gelfond.md): Under the per-digit certificate (W), which holds at every one-missing-digit base from 92317, twisting a missing-digit design with two consecutive digits by its digit sum costs its transform a fixed share of its mass at every frequency once norm((b - 1) alpha) > 0, so the major arcs of the unconditional dissection need no L-function, the Mobius and von Mangoldt sums over the design save a power of x, and the primes of the design equidistribute in the classes of their digit sum mod q coprime to b - 1; with the design read in progressions mod every N coprime to b, the digit sum of its primes is uniform mod every q on the classes prime to gcd(q, b - 1). - [The hexagon moire](https://mrly.net/raw/research/notes/hexagon.md): The moire of the stacked diagonal slices: the exact cut-ink laws, the doubling sign law, the quarter-line bands, and the ghost star, whose cell-frame decay is a closed form at every band width, -1/4 at the arm and -1/8 in the limit, every layer-pair constant an exact rational, and no frame-free coefficient. - [Information](https://mrly.net/raw/research/notes/information.md): A render as data: the rearranged SVD reads a design's code out of noise and peels a magic word, while the fractal codebook loses to deflate. - [Integers](https://mrly.net/raw/research/notes/integers.md): The census the other way round: which integers the whole registry writes, which it never writes, and which it writes thousands of times. - [Mobius and primes on the Kummer sets](https://mrly.net/raw/research/notes/kummer.md): At a prime p >= 5 the integers k with p not dividing C(3k, k) form a two-state carry automaton and not a digit design; at every prime p > a the Kummer set {k : p does not divide C(ak, bk)} keeps exactly ((p+1)/2)^L of the first p^L integers and its transform is a product of b(a - b)-square matrices of digit kernels, 2 x 2 at (3, 1); there a matrix transfer certificate clears 1/5 from p = 13567831, and from there the Mobius function cancels on the set and its primes are counted with constant p/(p+1). - [Magic words](https://mrly.net/raw/research/notes/magic.md): A different design at every scale: the word grammar, what collapses by block reduction, and what only the letter order can see. - [Method](https://mrly.net/raw/research/notes/method.md): How the results here are produced and checked, worked through on the odd-side fill polynomial. - [The Mobius meter across digit designs](https://mrly.net/raw/research/notes/mobius.md): The Mobius meter across digit designs: exact transfer between scaled columns, the 47-column cancellation census, a power saving on the dense columns under GRH, the pair route with its major-arc lemma and its l^1 threshold 1/4, and square-root cancellation as the open exponent. - [The fractal mirror](https://mrly.net/raw/research/notes/multilayers.md): Light through a stack of glass laid on a design: the level recursion that makes b^k layers cost (b - 1)k multiplications, the Born drift m(P_D) of the digit polynomial, the fixed-contrast step law that reads the two-variable lift Q_D instead, the cyclotomic designs that still go dark, the deep drift that misses m(Q_D) at weak contrast, and the conjectured split between critical stacks, whose passband shrinks as 1/sqrt(k), and the rest, which darken geometrically. - [Pi out of the stack](https://mrly.net/raw/research/notes/pi.md): Pi recovered three disjoint ways: the density of coprime points in the stacked set, one design's fill fluctuation, and one design across rising odd sides. - [REFS](https://mrly.net/raw/research/REFS.md): Every named reference, sequence id, theorem and attribution on this tree's root pages, resolved to a canonical URL. Ids that appear only inside a lab study are resolved in that study's own pages. - [Sequences](https://mrly.net/raw/research/sequences.md): Integer sequences fall out of the fractal work: cell counts, coprimality counts, Euler characteristics, design counts. This page is the ledger they are cited from. - [Slices](https://mrly.net/raw/research/notes/slices.md): The diagonal slice of the solid cube: the 6n census, centered-hexagonal vertices, and the splitting-prime rule. - [Spectra](https://mrly.net/raw/research/notes/spectra.md): The tile grammar of the diagonal slice at every odd base: the two-tile claim, the closed forms, and the mod-4 split. - [Spin](https://mrly.net/raw/research/notes/spin.md): A design turned about its centre: the ripple identity, the complete spin spectrum, the sponge's opaque diagonal, and a Gaussian Farey. - [The spirograph](https://mrly.net/raw/research/notes/spirograph.md): The roulette a design draws: the coincidence law that counts its distinct curves, the node count 2ab C(k,2) + k a(b-1) inside the window min(1, A) with the plane cut into nodes + 2 regions, the exact crossing staircase above the loop threshold whose thresholds are algebraic numbers obeying a reciprocity, the alignment reaches where four of the carpet's curves run through one point, and the walls, whose cover no closed form tried survives. - [The algebra of the stack](https://mrly.net/raw/research/notes/stack.md): The algebra of the stack: stacking is Dirichlet convolution and the group knows nothing new, the exact spun stack is a Gaussian Farey whose Fourier L^2 discrepancy obeys Franel's identity one field up and is equivalent to the Riemann hypothesis for zeta(s) L(s, chi_-4), its hexagonal twin puts L(2, chi_-3) in a proved node constant, the layers are a dilation system whose symbol is zeta(1 + s) so no reweighting moves the line, and the lean is the operation outside the Dirichlet group that still lands a closed form. - [Which designs tile](https://mrly.net/raw/research/notes/tilings.md): Which levels of a dimension-one design tile the integers: a closed condition on the zero set of the digit transform decides T1 at every level, a horizon decides T2, the guess that the digit set must tile its residues fails at base 12, a level fails T2 first at base 18, spectral levels agree with the tiling verdict on every case the census decides, and a design whose every level tiles has a spectral limit measure at prime-power size, at pq digits, at four digits or fewer, and on every case of the census, while a spectral limit measure whose levels are sets needs integer orthogonal sets of every size |F|^L, which proves the converse at five digits with 0 in F and levels that are sets and decides it on every set of its prime-size census. - [The toolpath](https://mrly.net/raw/research/notes/toolpaths.md): Self-similar curves on the triangular lattice as one-stroke print paths, and the sharp turn a printer slows for: the edge-covering kind never turns 60 degrees, every point-avoiding replacement curve of order 3 to 13 (order 12 without mirrors) turns 120 degrees at some level, the junction law that makes that a finite check, the exact sharp-turn count of the Gosper curve, the bead axis, which on the Gosper curve fades as 7^(-k/2), slower than Hilbert's, while Peano's holds at 1/2, and the radix census, where unit steps at every level are one identity at level 1 and a short periodic check and no plane-filling replacement curve with every flag F to a chord among the seven radix bases avoids points. - [The walk dimension](https://mrly.net/raw/research/notes/walks.md): The walk-dimension census: same mass, different music. - [Weighted designs](https://mrly.net/raw/research/notes/weights.md): Weighted designs: what weights move on the mass side, what they never move on the length side, and the pressure that closes both. - [Windows of a design](https://mrly.net/raw/research/notes/windows.md): The infinite design of a plane code as a subshift: its k x k windows counted exactly through the substitution, the carpet's count, conjecturally piecewise quadratic with p(k)/k^2 of lower limit 8 and upper limit 9.6, every code at bases 2 and 3 of growth 1, k or k^2, of finite type only when empty, full or on the boundary, and sofic always. - [The zeta function of a digit design](https://mrly.net/raw/research/notes/zeta.md): The design's own zeta function: an explicit zero-free half plane and the census it closes, a residue comb whose teeth are Rouche-certified one by one, a second family that is no critical line and obeys no counting, symmetry, contraction or gain law, the ordinate shadow as a constant-free Newton step from each zeta zero, the multiplicativity wall with the three products that stand where the Euler product does not, and the identity zeta_F N_F = 1 whose Mertens function runs the wrong way. ## Claims - [Acoustic barrier](https://mrly.net/raw/research/claims/acoustic-barrier.md): 1 claim, newest 2026-08-28 - [Adjacent digits](https://mrly.net/raw/research/claims/adjacent-digits.md): 11 claims, newest 2026-10-05 - [Arc loops](https://mrly.net/raw/research/claims/arc-loops.md): 32 claims, newest 2026-10-03 - [Atlas](https://mrly.net/raw/research/claims/atlas.md): 10 claims, newest 2026-09-08 - [Automata](https://mrly.net/raw/research/claims/automata.md): 20 claims, newest 2026-09-19 - [b-visibility](https://mrly.net/raw/research/claims/b-visibility.md): 2 claims, newest 2026-09-21 - [Boolean complexity](https://mrly.net/raw/research/claims/boolean-complexity.md): 2 claims, newest 2026-08-28 - [Carpet star](https://mrly.net/raw/research/claims/carpet-star.md): 2 claims, newest 2026-09-21 - [Cobham](https://mrly.net/raw/research/claims/cobham.md): 68 claims, newest 2026-09-23 - [Complex dimensions](https://mrly.net/raw/research/claims/complex-dimensions.md): 4 claims, newest 2026-09-06 - [Component counts](https://mrly.net/raw/research/claims/component-counts.md): 1 claim, newest 2026-08-28 - [Conduction at large side](https://mrly.net/raw/research/claims/conduction-at-large-side.md): 11 claims, newest 2026-10-03 - [Conjecture S: even half](https://mrly.net/raw/research/claims/conjecture-s-even-half.md): 19 claims, newest 2026-08-28 - [Conjecture S: odd half](https://mrly.net/raw/research/claims/conjecture-s-odd-half.md): 3 claims, newest 2026-08-28 - [Conjecture S: the even-half transient](https://mrly.net/raw/research/claims/conjecture-s-the-even-half-transient.md): 3 claims, newest 2026-08-28 - [Coprimality at dimension one](https://mrly.net/raw/research/claims/coprimality-at-dimension-one.md): 12 claims, newest 2026-09-21 - [Coprimality density](https://mrly.net/raw/research/claims/coprimality-density.md): 16 claims, newest 2026-09-21 - [Crop census](https://mrly.net/raw/research/claims/crop-census.md): 20 claims, newest 2026-09-06 - [Density theorem boundary](https://mrly.net/raw/research/claims/density-theorem-boundary.md): 13 claims, newest 2026-09-21 - [Dependent bases](https://mrly.net/raw/research/claims/dependent-bases.md): 8 claims, newest 2026-09-14 - [Design census](https://mrly.net/raw/research/claims/design-census.md): 6 claims, newest 2026-08-28 - [Design counts at every base](https://mrly.net/raw/research/claims/design-counts-at-every-base.md): 1 claim, newest 2026-08-28 - [Design windows](https://mrly.net/raw/research/claims/design-windows.md): 19 claims, newest 2026-10-03 - [Diagonal cuts of the parity solid](https://mrly.net/raw/research/claims/diagonal-cuts-of-the-parity-solid.md): 2 claims, newest 2026-08-28 - [Diagonal designs](https://mrly.net/raw/research/claims/diagonal-designs.md): 1 claim, newest 2026-08-28 - [Diagonal slice ladder](https://mrly.net/raw/research/claims/diagonal-slice-ladder.md): 22 claims, newest 2026-09-23 - [Diagonal slice stack](https://mrly.net/raw/research/claims/diagonal-slice-stack.md): 30 claims, newest 2026-09-21 - [Digit designs and the Euler product](https://mrly.net/raw/research/claims/digit-designs-and-the-euler-product.md): 36 claims, newest 2026-09-21 - [Digit strings across divisors](https://mrly.net/raw/research/claims/digit-strings-across-divisors.md): 21 claims, newest 2026-09-19 - [Dilated receptive fields](https://mrly.net/raw/research/claims/dilated-receptive-fields.md): 18 claims, newest 2026-09-23 - [Dimension one](https://mrly.net/raw/research/claims/dimension-one.md): 26 claims, newest 2026-09-21 - [Dimers on a design](https://mrly.net/raw/research/claims/dimers-on-a-design.md): 24 claims, newest 2026-10-03 - [Divisor avatars](https://mrly.net/raw/research/claims/divisor-avatars.md): 3 claims, newest 2026-08-28 - [Erdos](https://mrly.net/raw/research/claims/erdos.md): 47 claims, newest 2026-10-05 - [Exposure](https://mrly.net/raw/research/claims/exposure.md): 1 claim, newest 2026-08-30 - [Farey stack](https://mrly.net/raw/research/claims/farey-stack.md): 17 claims, newest 2026-09-21 - [Fill polynomials](https://mrly.net/raw/research/claims/fill-polynomials.md): 38 claims, newest 2026-09-21 - [Flake band gap](https://mrly.net/raw/research/claims/flake-band-gap.md): 2 claims, newest 2026-08-28 - [Flat carpet stack](https://mrly.net/raw/research/claims/flat-carpet-stack.md): 16 claims, newest 2026-09-21 - [Franel on a digit design](https://mrly.net/raw/research/claims/franel-on-a-digit-design.md): 76 claims, newest 2026-10-05 - [Franel one field up](https://mrly.net/raw/research/claims/franel-one-field-up.md): 3 claims, newest 2026-09-09 - [Gasket rays and the window](https://mrly.net/raw/research/claims/gasket-rays-and-the-window.md): 30 claims, newest 2026-09-21 - [Gaussian primes on a design](https://mrly.net/raw/research/claims/gaussian-primes-on-a-design.md): 25 claims, newest 2026-10-05 - [Half interval](https://mrly.net/raw/research/claims/half-interval.md): 14 claims, newest 2026-10-03 - [Half-ball chords](https://mrly.net/raw/research/claims/half-ball-chords.md): 4 claims, newest 2026-08-28 - [Hexagram provenance](https://mrly.net/raw/research/claims/hexagram-provenance.md): 7 claims, newest 2026-09-21 - [Integer census and avatars](https://mrly.net/raw/research/claims/integer-census-and-avatars.md): 4 claims, newest 2026-08-28 - [Kronecker word order](https://mrly.net/raw/research/claims/kronecker-word-order.md): 16 claims, newest 2026-09-19 - [Kronecker words](https://mrly.net/raw/research/claims/kronecker-words.md): 8 claims, newest 2026-09-19 - [Kummer sets](https://mrly.net/raw/research/claims/kummer-sets.md): 10 claims, newest 2026-10-05 - [Level tilings](https://mrly.net/raw/research/claims/level-tilings.md): 63 claims, newest 2026-10-05 - [Levels and designs](https://mrly.net/raw/research/claims/levels-and-designs.md): 9 claims, newest 2026-09-09 - [Matched random control](https://mrly.net/raw/research/claims/matched-random-control.md): 1 claim, newest 2026-08-28 - [Memory zeta](https://mrly.net/raw/research/claims/memory-zeta.md): 1 claim, newest 2026-09-21 - [Menger slice sources](https://mrly.net/raw/research/claims/menger-slice-sources.md): 5 claims, newest 2026-08-28 - [Mertens stack](https://mrly.net/raw/research/claims/mertens-stack.md): 1 claim, newest 2026-08-28 - [Moire local limit](https://mrly.net/raw/research/claims/moire-local-limit.md): 16 claims, newest 2026-10-02 - [Moment ladder and Lemma B](https://mrly.net/raw/research/claims/moment-ladder-and-lemma-b.md): 5 claims, newest 2026-08-31 - [Multilayers](https://mrly.net/raw/research/claims/multilayers.md): 12 claims, newest 2026-09-23 - [Odd-base slice grammar](https://mrly.net/raw/research/claims/odd-base-slice-grammar.md): 19 claims, newest 2026-09-23 - [Odd-side fills](https://mrly.net/raw/research/claims/odd-side-fills.md): 1 claim, newest 2026-08-30 - [One number at every odd side](https://mrly.net/raw/research/claims/one-number-at-every-odd-side.md): 17 claims, newest 2026-10-02 - [Primes on a design](https://mrly.net/raw/research/claims/primes-on-a-design.md): 40 claims, newest 2026-09-23 - [Robin corridor](https://mrly.net/raw/research/claims/robin-corridor.md): 1 claim, newest 2026-08-28 - [Shortest paths](https://mrly.net/raw/research/claims/shortest-paths.md): 15 claims, newest 2026-10-02 - [Slice at large side](https://mrly.net/raw/research/claims/slice-at-large-side.md): 11 claims, newest 2026-10-02 - [Slice sign law in every dimension](https://mrly.net/raw/research/claims/slice-sign-law-in-every-dimension.md): 5 claims, newest 2026-08-28 - [Sparse arrays](https://mrly.net/raw/research/claims/sparse-arrays.md): 11 claims, newest 2026-10-05 - [Spectral spacings](https://mrly.net/raw/research/claims/spectral-spacings.md): 1 claim, newest 2026-08-28 - [Spin](https://mrly.net/raw/research/claims/spin.md): 31 claims, newest 2026-09-21 - [Sponge measurability](https://mrly.net/raw/research/claims/sponge-measurability.md): 12 claims, newest 2026-09-23 - [Stacked hexagon moire](https://mrly.net/raw/research/claims/stacked-hexagon-moire.md): 6 claims, newest 2026-09-21 - [The 2-adic Smith cascade](https://mrly.net/raw/research/claims/2-adic-smith-cascade.md): 29 claims, newest 2026-09-14 - [The Apollonian gasket](https://mrly.net/raw/research/claims/apollonian-gasket.md): 22 claims, newest 2026-09-21 - [The arithmetic pole](https://mrly.net/raw/research/claims/arithmetic-pole.md): 1 claim, newest 2026-09-06 - [The circle count](https://mrly.net/raw/research/claims/circle-count.md): 17 claims, newest 2026-09-19 - [The cocycle's joint spectral radius](https://mrly.net/raw/research/claims/cocycle-joint-spectral-radius.md): 2 claims, newest 2026-09-06 - [The codebook](https://mrly.net/raw/research/claims/codebook.md): 3 claims, newest 2026-09-06 - [The coin](https://mrly.net/raw/research/claims/coin.md): 11 claims, newest 2026-10-03 - [The Collatz carry](https://mrly.net/raw/research/claims/collatz-carry.md): 10 claims, newest 2026-09-14 - [The component exponent](https://mrly.net/raw/research/claims/component-exponent.md): 22 claims, newest 2026-08-31 - [The design horocycle](https://mrly.net/raw/research/claims/design-horocycle.md): 11 claims, newest 2026-10-05 - [The design Mobius meter](https://mrly.net/raw/research/claims/design-mobius-meter.md): 4 claims, newest 2026-09-19 - [The diagonal cut](https://mrly.net/raw/research/claims/diagonal-cut.md): 2 claims, newest 2026-09-21 - [The digit progressions](https://mrly.net/raw/research/claims/digit-progressions.md): 13 claims, newest 2026-10-05 - [The digit-restricted Mobius exponent](https://mrly.net/raw/research/claims/digit-restricted-mobius-exponent.md): 18 claims, newest 2026-09-23 - [The digit-restricted Mobius meter](https://mrly.net/raw/research/claims/digit-restricted-mobius-meter.md): 65 claims, newest 2026-09-19 - [The digit-sum twist](https://mrly.net/raw/research/claims/digit-sum-twist.md): 20 claims, newest 2026-10-05 - [The Eisenstein stack](https://mrly.net/raw/research/claims/eisenstein-stack.md): 2 claims, newest 2026-09-09 - [The even digits](https://mrly.net/raw/research/claims/even-digits.md): 7 claims, newest 2026-10-05 - [The hexagonal slice](https://mrly.net/raw/research/claims/hexagonal-slice.md): 3 claims, newest 2026-08-28 - [The layers as a dilation system](https://mrly.net/raw/research/claims/layers-as-a-dilation-system.md): 4 claims, newest 2026-09-09 - [The leaning stack](https://mrly.net/raw/research/claims/leaning-stack.md): 4 claims, newest 2026-09-09 - [The memory dial](https://mrly.net/raw/research/claims/memory-dial.md): 100 claims, newest 2026-09-21 - [The moments of the digit transform](https://mrly.net/raw/research/claims/moments-of-the-digit-transform.md): 5 claims, newest 2026-09-07 - [The node stack](https://mrly.net/raw/research/claims/node-stack.md): 3 claims, newest 2026-09-09 - [The novelty meter](https://mrly.net/raw/research/claims/novelty-meter.md): 10 claims, newest 2026-09-21 - [The parity fill](https://mrly.net/raw/research/claims/parity-fill.md): 4 claims, newest 2026-09-09 - [The parity-carpet stack spectrum](https://mrly.net/raw/research/claims/parity-carpet-stack-spectrum.md): 1 claim, newest 2026-08-28 - [The question mark](https://mrly.net/raw/research/claims/question-mark.md): 33 claims, newest 2026-09-23 - [The radix census](https://mrly.net/raw/research/claims/radix-census.md): 20 claims, newest 2026-10-03 - [The radix dial](https://mrly.net/raw/research/claims/radix-dial.md): 42 claims, newest 2026-09-14 - [The ratio-set power saving](https://mrly.net/raw/research/claims/ratio-set-power-saving.md): 30 claims, newest 2026-09-19 - [The registry's integers](https://mrly.net/raw/research/claims/registry-integers.md): 23 claims, newest 2026-08-31 - [The repunit layer](https://mrly.net/raw/research/claims/repunit-layer.md): 9 claims, newest 2026-09-07 - [The second moment of the rays](https://mrly.net/raw/research/claims/second-moment-of-the-rays.md): 2 claims, newest 2026-08-31 - [The slice ladder: rate and dead routes](https://mrly.net/raw/research/claims/slice-ladder-rate-and-dead-routes.md): 11 claims, newest 2026-08-28 - [The spirograph loops](https://mrly.net/raw/research/claims/spirograph-loops.md): 25 claims, newest 2026-09-10 - [The spirograph nodes](https://mrly.net/raw/research/claims/spirograph-nodes.md): 7 claims, newest 2026-09-10 - [The spirograph reaches](https://mrly.net/raw/research/claims/spirograph-reaches.md): 27 claims, newest 2026-09-10 - [The spirograph walls](https://mrly.net/raw/research/claims/spirograph-walls.md): 12 claims, newest 2026-09-10 - [The spun stack](https://mrly.net/raw/research/claims/spun-stack.md): 12 claims, newest 2026-09-09 - [The stack algebra](https://mrly.net/raw/research/claims/stack-algebra.md): 5 claims, newest 2026-09-09 - [The staircase law](https://mrly.net/raw/research/claims/staircase-law.md): 9 claims, newest 2026-10-02 - [The tent rank law](https://mrly.net/raw/research/claims/tent-rank-law.md): 4 claims, newest 2026-08-28 - [The tile monoid](https://mrly.net/raw/research/claims/tile-monoid.md): 23 claims, newest 2026-08-31 - [The unequal split](https://mrly.net/raw/research/claims/unequal-split.md): 18 claims, newest 2026-10-03 - [The zeros of the design zeta](https://mrly.net/raw/research/claims/zeros-of-the-design-zeta.md): 26 claims, newest 2026-09-19 - [Toolpaths](https://mrly.net/raw/research/claims/toolpaths.md): 9 claims, newest 2026-09-23 - [Transparent fraction](https://mrly.net/raw/research/claims/transparent-fraction.md): 2 claims, newest 2026-09-21 - [Walk dimension](https://mrly.net/raw/research/claims/walk-dimension.md): 2 claims, newest 2026-08-28 - [Wallis sieve](https://mrly.net/raw/research/claims/wallis-sieve.md): 4 claims, newest 2026-09-06 - [Waves](https://mrly.net/raw/research/claims/waves.md): 14 claims, newest 2026-09-19 - [Weighted designs](https://mrly.net/raw/research/claims/weighted-designs.md): 4 claims, newest 2026-09-06 ## Papers - [A Power Saving for the Digit Sum on Missing Digits](https://mrly.net/raw/research/papers/digit-sum-twist-power-saving.md): On the integers whose base-b digits avoid a fixed set, keeping two consecutive digits and the one-step certificate (W), the Mobius and von Mangoldt functions modulated by the digit sum, e(alpha s_b(n)), save a power of x against the count of the set whenever (b - 1) alpha is not an integer, with an explicit exponent, an effective constant and no L-function; (W) holds at every one-missing-digit set from base 92317, and the primes of the set then equidistribute by digit sum modulo every q coprime to b - 1. The power is tiny, delta_0 >= 4.678 * 10^-9 at 92317 for every alpha with norm((b - 1) alpha) >= 0.2: the point is the mechanism. - [The Exact Fragility of Fractal Sparse Arrays](https://mrly.net/raw/research/papers/fractal-array-fragility.md): From order 2 on, a sensor of a fractal sparse array is essential exactly when every base-M digit of its position has a partner in the generator at a lag that no other pair makes, so the fragility is exactly (u/L)^r; from order 2 on, the bound of Cohen and Eldar is attained exactly when every sensor of the generator has such a partner, which is the strong converse of their Theorem 5, and the power bound of Yang and coauthors exactly when every essential sensor does. - [Spectral Measures from Tiling Levels](https://mrly.net/raw/research/papers/spectral-tiling-levels.md): For an integer base and a digit set whose every level tiles the integers, the self-similar measure is spectral when the number of digits is a prime power or a product of two primes; at base 12 the digits {0, 1, 8, 9} give such a measure with an explicit spectrum though they are no multiple of a complementing set; and at five digits containing 0, with levels that are sets, the measure is spectral if and only if every level tiles, so {0, 1, 3, 5, 6} gives a spectral measure at no integer base. - [An Unconditional Mertens Bound on the Half Interval](https://mrly.net/raw/research/papers/half-interval-mobius.md): At every odd base base >= 94939, on the integers whose digits all lie in the lower half {0, ..., (base-1)/2}, the Mobius function cancels, |M_F(x)| <= C A_F(x) exp(-c sqrt(log x)), and the primes are counted, the von Mangoldt sum being (base/(base+1)) A_F(x) within the same error, both proved here with no hypothesis and computable constants; at a prime base p that set is {k : p does not divide C(2k,k)}, and a transfer operator whose cost per digit is (2/pi) log base + O(1) sets the base. - [A Moving Target for Erdos Problem 125](https://mrly.net/raw/research/papers/erdos-125-upper-density.md): The sums of a base-3 number and a base-4 number, each written with the digits 0 and 1 only, have upper density at least 6/(11 liminf_k h_k(tau_k)), where h_k is the cell energy of a Cantor convolution read at the scaling tau_k = 4^(m_k)/3^k; so the open half of Erdos problem 125 fails only if one orbit of an irrational rotation lands, at every late step, in a target of measure at most 190/M, for every M. - [An Unconditional Mertens Bound at Large Base](https://mrly.net/raw/research/papers/unconditional-mertens-at-large-base.md): The Mobius function summed over the integers that avoid one digit in base q is at most C A_F(x) exp(-c sqrt(log x)) with no hypothesis at every base q >= 584 by a proof, and at every base q >= 115 on certificates computed base by base, with C and c effective; Dirichlet approximation cuts the frequencies into four regions, and only the minor arcs set the base. - [The First Base Below a Quarter](https://mrly.net/raw/research/papers/first-base-below-a-quarter.md): Base 21 missing the digit 0 is the least base at which a one-missing-digit set has the l^1 exponent of its digit transform below the bar 1/4, all 108 sets of the bases 3 to 20 certified above it; 34 is the least base at which every digit clears and every base from 34 clears at every digit, from 125 by a proof; under the generalized Riemann hypothesis each clearance is a power saving for the Mobius sum over the set. - [The Franel Threshold Is a Mean Square](https://mrly.net/raw/research/papers/franel-converse.md): On the Farey fractions whose denominators lie in a digit design, Franel's square-root threshold is equivalent to square-root cancellation, in mean square over the dilates of the design, of their Mertens sums, with explicit logarithms in the sandwich; the full set recovers Franel. - [The Stack Hears the Zeros](https://mrly.net/raw/research/papers/novelty-meter.md): The novelty of the stack, read through a smooth window, has an error that is a sum of waves at the frequencies of the zeros of zeta; its decay exponent is a Riemann hypothesis meter, and a sharp window hears the primes instead. - [The Menger Sponge Is Not Minkowski Measurable](https://mrly.net/raw/research/papers/sponge-measurability.md): The Menger sponge is not Minkowski measurable, by the criterion of Kombrink, Pearse and Winter with the open unit cube as feasible set, which needs no pluriphase hypothesis, and an interval-arithmetic certificate that puts its periodic function at two phases at least 0.012296 apart. - [The Spin Spectrum Reads a Pair Census](https://mrly.net/raw/research/papers/spin-harmonics.md): The spin spectrum of a design is one fixed linear function of its pair census, so some designs are homometric and no spin reading one level of refinement can tell them apart. - [Coprimality Density Above Dimension One](https://mrly.net/research/papers/coprime-density-above-dimension-one/): the visible-point density of every digit-restricted fractal above dimension one, exactly. - [The Pincer at Dimension One: Two Edges and Two Closed Doors](https://mrly.net/research/papers/lemma-b-pincer/): the missing dimension-one estimate squeezed from both edges, and two routes past it proved shut. - [A Power Saving for Mobius Sums on Missing-Digit Integers, under GRH](https://mrly.net/research/papers/sparse-mertens-under-grh/): under GRH, the Mobius function cancels by a power over the integers missing a digit, at every base from 3690 up. - [Pairwise Coprimality in the Menger Sponge](https://mrly.net/research/papers/menger-pairwise-coprimality/): drilling the sponge's holes costs its coordinates exactly 12.25 percent of their pairwise-coprimality odds. - [The Walsh Spectrometer: Exact Diagonal-Slice Ink for Every Three-Dimensional Parity Design](https://mrly.net/research/papers/walsh-spectrometer/): the exact slice ink of all 256 parity designs, with the design's Walsh spectrum as the coefficients. - [Menger Diagonal Slices: A Recurrence of Order ⌈D/2⌉](https://mrly.net/research/papers/slice-recurrence-order/): Menger diagonal slice counts obey a recurrence of order exactly ceil(D/2), and at odd D != 1 mod 3 their exponent provably beats the generic slice dimension. - [The Even Half of the Slice Sign Law](https://mrly.net/research/papers/slice-sign-even-half/): the even half of the slice sign law: below the generic exponent in every even dimension at bases 3 and 5, the quadratic base-3 transient identified exactly, and the Jacobsthal tent rank law behind the remaining strictness gap. - [Base-3 Digit Designs: Diagonality, Ray Masses, and a Spectral Gap at Two](https://mrly.net/research/papers/gasket-ray-machine/): which origin lines hit a base-3 design, how much each catches, and a spectral gap at two. - [Order Sensitivity of Kronecker Design Words: What the Perfect Shuffle Cannot See](https://mrly.net/research/papers/order-sensitivity-of-kronecker-words/): what survives swapping nested patterns, and the connectivity that does not. - [The Component Exponent of a Two-Letter Kronecker Word](https://mrly.net/research/papers/component-exponent-of-kronecker-words/): the piece count of a two-letter nesting has a closed form on all 105 alphabets, so its growth rate exists, is order-blind, and mostly just repeats the cell count. - [Parity Carpets Correlate by gcd: Four Exact Laws for a Square-Wave Stack](https://mrly.net/research/papers/moire-correlation-laws/): two parity carpets agree by an exact gcd law, zero exactly when the scales are coprime; the stack's full spectrum is the squared-divisor field of the frequency gcd, zeta quotients and nothing more. - [Divisor Avatars: Which Parity Designs Count the Divisors of a Power](https://mrly.net/research/papers/divisor-avatars/): when a parity design's cell census is the divisor count of a power, and exactly which integers have one. - [The Sequence Census of a Parity Design](https://mrly.net/research/papers/sequence-census/): how many integer sequences a parity rule can write, which ones the catalogue already holds, and why the polygonal numbers keep turning up. - [The Dirichlet Inverse of a Digit Design and the Transport of Zeros](https://mrly.net/research/papers/design-dirichlet-inverse/): the design's own Mobius function, and the zeros of its zeta that push its partial sums past the design's own size; still owes its outside read. ## Blog - [Launching the site](https://mrly.net/raw/site/blog/launching-mrlyprod-org/index.md): One address for the demos, the papers and the research notes. ## Math - [Math](https://mrly.net/raw/pkgs/mrlyrs/NAMES.md): Words, not letters. One word per thing, the same word everywhere. Letters live only inside a formula, declared by the sentence that uses them. Implemented by mrlyrs::math::name behind the Named trait. ## About - [About](https://mrly.net/raw/site/pages/about.md): One written mark, the mathematics under it, and the code that draws it. - [Contact](https://mrly.net/raw/site/pages/contact.md): Where to write, and every account this company keeps. - [Donate](https://mrly.net/raw/site/pages/donate.md): Everything here is free; a donation is what keeps it that way. - [Method](https://mrly.net/raw/site/pages/method.md): Carlo directs, Claude writes and computes, and every number names its generator.