THE FOLD / CO-OP / THE SYNC / THE RUDIN-SHAPIRO
THE RUDIN-SHAPIRO
a deterministic +-1 sequence with random-walk-flat sums
1 WHAT IT IS · WHAT IT DOES · FACT OR FICTION
The Rudin–Shapiro sequence rₙ = (−1)(number of ‘11’ pairs in the binary of n) is a ±1 sequence engineered to be flat: its partial sums stay astonishingly small — growing like √N rather than N — and equivalently its power spectrum is nearly flat (very low autocorrelation).
That flatness is prized: spreading a signal’s energy evenly across a band is exactly what radar pulse-compression and spread-spectrum communication need, and low-correlation ±1 sequences like this one make it possible.
LIT verified live: over N up to 200,000, the ratio |SN|/√N of the partial sums stays bounded (measured max ≈ 2.45, within the proven bound 2+√2 ≈ 3.41), while a structureless sum would grow linearly (window.__rudinshapiro). FIG no framing; exact bit-counting.
That flatness is prized: spreading a signal’s energy evenly across a band is exactly what radar pulse-compression and spread-spectrum communication need, and low-correlation ±1 sequences like this one make it possible.
LIT verified live: over N up to 200,000, the ratio |SN|/√N of the partial sums stays bounded (measured max ≈ 2.45, within the proven bound 2+√2 ≈ 3.41), while a structureless sum would grow linearly (window.__rudinshapiro). FIG no framing; exact bit-counting.
2 HOW IT WAS WEAVED · AI + HUMAN
David (human) seated this at the-sync — correlation and synchronisation, where flat spectra let receivers lock on. The Rudin–Shapiro sequence is the sync engineer’s friend: deterministic, yet noise-flat. AVAN (AI) built the instrument: the bit-pair parity rule, the partial-sum walk, the √N envelope check.
Credit as content: Harold Shapiro (1951 thesis) and Walter Rudin (1959); these are ‘Golay–Rudin–Shapiro’ sequences in signal processing. The weave: David names the sync; I generate the ±1 sequence from binary bit-pairs and show its sums hug a √N envelope no random-looking sum should respect so tightly.
Credit as content: Harold Shapiro (1951 thesis) and Walter Rudin (1959); these are ‘Golay–Rudin–Shapiro’ sequences in signal processing. The weave: David names the sync; I generate the ±1 sequence from binary bit-pairs and show its sums hug a √N envelope no random-looking sum should respect so tightly.
3 ONE DIMENSION
The ±1 sequence (from the parity of ‘11’ bit-pairs) and the running partial sum — a walk that, unlike most deterministic sums, never drifts far from zero.
4 TWO DIMENSIONS · INTERACTIVE
Plot the partial sums SN against the ±C√N envelope. Extend N and watch the walk stay inside the square-root band, its peak ratio |SN|/√N holding near 2.45.
5 THREE DIMENSIONS + AVAN’S INVERSE
The green forward object: the Rudin–Shapiro partial-sum walk, hugging the √N envelope as it grows.
AVAN’s addition (the inverse-companion): a fully deterministic sequence behaves, where it counts, like a random one. Its partial sums grow like √N — exactly the scale of a random walk of ±1 coin flips — and its spectrum is flat like white noise, yet every term is fixed by a two-line binary rule. The inverse of ‘structured and predictable’ is ‘as balanced as coin flips, on purpose.’ This is designed flatness: it is the mirror of Thue–Morse, whose rule designs imbalance-avoidance — here the rule designs correlation-avoidance, producing noise-like statistics from rigid structure. Magenta is the √N random-walk envelope; green is the deterministic walk that never escapes it. Rigidity engineered to look like chance.
LIT Genuine Rudin-Shapiro (Golay-Rudin-Shapiro) sequence (Shapiro 1951; Rudin 1959). Verified live: computing r_n from the parity of '11' bit-pairs, the partial-sum ratio |S_N|/sqrt(N) over N<=200000 has measured maximum ~2.45 (at N=174763), within the proven bound 2+sqrt2~3.414 — bounded, unlike a linearly-growing structureless sum (window.__rudinshapiro.bounded).
FIG No framing: the bit-pair parity rule and the partial-sum/envelope check run in-browser and are exact. The AVAN inverse is honest — the sums genuinely grow at the sqrt(N) random-walk scale and the spectrum is flat like noise despite a rigid rule; it is the designed-correlation-avoidance mirror of Thue-Morse's designed-imbalance-avoidance. Magenta is the sqrt(N) envelope, green the deterministic walk.
FIG No framing: the bit-pair parity rule and the partial-sum/envelope check run in-browser and are exact. The AVAN inverse is honest — the sums genuinely grow at the sqrt(N) random-walk scale and the spectrum is flat like noise despite a rigid rule; it is the designed-correlation-avoidance mirror of Thue-Morse's designed-imbalance-avoidance. Magenta is the sqrt(N) envelope, green the deterministic walk.
◆ sealed .dlw.fold → folded to ROOT_0 · a sphere of THE SYNC · David Lee Wise (ROOT0), with AVAN