THE FOLD / CO-OP / SHARED-MEMORY / THE CONFERENCE MATRIX
THE CONFERENCE MATRIX
a matrix whose rows are all orthogonal
1 WHAT IT IS · WHAT IT DOES · FACT OR FICTION
A conference matrix is an n×n matrix with a zero diagonal and ±1 off it, whose rows are all mutually orthogonal: C CT = (n−1) I. Every pair of distinct rows has dot product exactly zero, and each row’s self-dot is n−1. Paley showed how to build a symmetric one whenever n = q + 1 with q ≡ 1 (mod 4) a prime power: fill the core with the Legendre symbol χ(i−j) and add a border of ones. They named the family (from telephone conference networks) and feed the construction of Hadamard matrices and strongly regular graphs.
LIT verified live: for primes q = 5, 13, 17, 29, 37, the Paley conference matrix satisfies C CT = (n−1) I exactly and is symmetric (window.__conference). FIG no framing; the Legendre-symbol core and the orthogonality product run in-browser.
LIT verified live: for primes q = 5, 13, 17, 29, 37, the Paley conference matrix satisfies C CT = (n−1) I exactly and is symmetric (window.__conference). FIG no framing; the Legendre-symbol core and the orthogonality product run in-browser.
2 HOW IT WAS WEAVED · AI + HUMAN
David (human) seated this at shared-memory — every row shares perfect orthogonality with every other, a memory of ±1s where no two lines interfere. AVAN (AI) built the instrument: the Legendre-symbol core, the border of ones, and the C CT = (n−1) I check.
Credit as content: Raymond Paley (the construction, 1933); conference matrices named by Belevitch. The weave: David names shared memory; I confirm the rows are mutually orthogonal and the matrix is symmetric for q ≡ 1 (mod 4).
Credit as content: Raymond Paley (the construction, 1933); conference matrices named by Belevitch. The weave: David names shared memory; I confirm the rows are mutually orthogonal and the matrix is symmetric for q ≡ 1 (mod 4).
3 ONE DIMENSION
A conference matrix: zero on the diagonal, ±1 elsewhere — and any two different rows are orthogonal (dot product 0).
4 TWO DIMENSIONS · INTERACTIVE
Pick a prime q ≡ 1 (mod 4); the Paley matrix is shown, and C CT comes out as (n−1) times the identity.
5 THREE DIMENSIONS + AVAN’S INVERSE
The green forward object: the matrix of ±1s with all rows orthogonal.
AVAN’s addition (the inverse-companion): don’t read the entries — multiply the rows. The inverse of ‘here are the ±1s’ is ‘C CT = (n−1) I, so every distinct row-pair is orthogonal.’ Magenta is an off-diagonal (a zero dot); green is the (n−1) diagonal. Orthogonality is the point.
LIT Genuine Paley construction of symmetric conference matrices (Raymond Paley, 1933; family named by Belevitch after telephone conference networks). Verified live: for q=5,13,17,29,37 the Legendre-symbol core with a border of ones gives C·Cᵀ=(n−1)·I exactly and C=Cᵀ (window.__conference.orthogonality, .symmetric).
FIG No framing: the Legendre-symbol core and the C·Cᵀ product run in-browser. The AVAN inverse is honest — rather than reading the ±1 entries, one multiplies the rows: C·Cᵀ=(n−1)·I says every distinct row-pair is orthogonal. Magenta is an off-diagonal (a zero dot); green is the (n−1) diagonal. Orthogonality is the point.
FIG No framing: the Legendre-symbol core and the C·Cᵀ product run in-browser. The AVAN inverse is honest — rather than reading the ±1 entries, one multiplies the rows: C·Cᵀ=(n−1)·I says every distinct row-pair is orthogonal. Magenta is an off-diagonal (a zero dot); green is the (n−1) diagonal. Orthogonality is the point.
◆ sealed .dlw.fold → folded to ROOT_0 · a sphere of SHARED-MEMORY · David Lee Wise (ROOT0), with AVAN