--- slug: every-pair-once passed: 2026-10-18 by: the assay line (research/assay-line/claude-1018b/), Claude fixer (claude-opus-5-5) directed by claude-assaying-boyle-1014; source record by a Claude agent (claude-sonnet-5-5), checked by a Claude agent (claude-opus-5-5) checked: 114 confirmed: 92 # 74 CONFIRMED + 18 MINOR (near-misses, counted here) wrong: 15 unverifiable: 3 observed: 4 stale: 0 --- This was the assay line's directed correction pass on /strata/every-pair-once/ (batch claude-1018b, tier B, read at db2da072f8), following research/assay-line/claude-1018b/every-pair-once/triage.json. Every one of the 33 triage items was decided before any edit, and every WRONG item's evidence was re-read by the fixer (Wikipedia's Kirkman, social golfer, Ronald Fisher and mutually orthogonal Latin squares articles; MathWorld's Social Golfer Problem; csplib prob010; Crossref for the Sylvester citation; the layer's verify.mjs and the every-difference-once markdown; Aguado's ten-week table re-checked in scratch, 480 distinct pairs). Fifteen WRONG lines (seven distinct errors) were fixed, each distinct error with one dated Corrected line on the page and in the markdown; four MINOR lines were fixed with no line (an altered quotation, and three that share a sentence with a WRONG fix); fourteen MINOR lines were declined under the tier B rule. The other 81 lines are carried forward from the Claude checker's record (N2-claude.md) as that checker read them, not as fresh reads by the fixer. JSON-LD dateModified was set to 2026-10-18. The verifier was re-run after the edits: 29/29. ## Claims - CONFIRMED | Kirkman posed this in 1850 in the Lady's and Gentleman's Diary | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [kirkman-posed-1850] "proposed by Thomas Penyngton Kirkman in 1850 as Query VI in The Lady's and Gentleman's Diary (pg.48)" - CONFIRMED | Query VI in The Lady's and Gentleman's Diary for the Year 1850 | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [query-vi-title] Query VI, 1850, p. 48 as the page says; the volume wording "for the Year 1850" is not in the source (the Diary itself could not be reached), the year and query number are - MINOR | Fifteen young ladies of a school walk out three abreast for seven days in succession; it is required to arrange them daily so that no two shall walk twice abreast. | https://en.wikipedia.org/wiki/Thomas_Kirkman | K1: QUOTATIONS rule: the page prints the 1850 query as Kirkman's wording with "of a school" and a semicolon; the Wikipedia article (raw wikitext, the lead blockquote) reads "Fifteen young ladies in a school walk out three abreast for seven days in succession: it is required...". The fix restores the source's words ("in", and the colon); one-word slip, so MINOR, no page line. FIXED: restored the 1850 wording in index.html ("in a school", colon), per Wikipedia's blockquote; MINOR, no page line. - CONFIRMED | Kirkman, T. P. (1850). Query VI. The Lady's and Gentleman's Diary for the Year 1850: 48. The original schoolgirl statement. | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [cit-kirkman-1850] "in 1850 as Query VI in The Lady's and Gentleman's Diary (pg.48)" - WRONG | because he had already shown it in a related setting three years earlier | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | K2: Re-read Wikipedia, Kirkman's schoolgirl problem (raw wikitext, History, the 1847 bullet): "Using these propositions, he proved that triple systems exist for all values of n = 1 or 3 (mod 6) (not necessarily resolvable ones...)" and "He could not conclusively say for what other values of n would resolvable triple systems exist". The 1847 paper showed the count IS enough for triple systems; it did not show a closing count failing. WRONG, fix. FIXED: index.html "because he had already shown it in a related setting three years earlier" replaced: in 1847 Kirkman proved that for plain triple systems the count is enough; Corrected line added. - MINOR | But books that close do not always keep, and Kirkman knew this | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | K3: MINOR in the same sentence as K2 (the clause "and Kirkman knew this" rests on the false 1847 clause); tier B rule: fixed with it, recast to what the 1847 paper did show. FIXED: same sentence as K2: "and Kirkman knew this" removed in the recast; MINOR, no line. - CONFIRMED | Kirkman, T. P. (1847). On a problem in combinations. Cambridge and Dublin Mathematical Journal 2: 191–204. | https://en.wikipedia.org/wiki/Steiner_system | [cit-kirkman-1847] "Kirkman, Thomas P. (1847), "On a Problem in Combinations", The Cambridge and Dublin Mathematical Journal, II: 191–204."; the IA OCR shows the paper's running heads on pp. 191 to 203 of vol. II, dated May 1847 - CONFIRMED | Existence of Steiner triple systems six years before Steiner's own query. | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [steiner-six-years] 1847 paper; "In 1853, Jakob Steiner, completely unaware of Kirkman's 1847 paper..."; 1853 - 1847 = 6 - CONFIRMED | Kirkman found one, and the deep reason it exists is a finite space. | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [kirkman-found-one] "Kirkman himself gave a solution that later would be found to be isomorphic to Solution I above" and "Both solutions could be embedded in PG(3,2)" - CONFIRMED | Kirkman showed the perfect answer exists. | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [kirkman-showed-exists] same passage: Kirkman gave a solution - CONFIRMED | Kirkman's 1850 construction shows the perfect answer | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [kirkman-1850-construction] same passage (md body, not rendered) - MINOR | Before there were fifteen there were nine. | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | K4: Tier B MINOR, not in a sentence being fixed: read as a narrative opener it is loose, and Wikipedia says the 1847 paper treated "n = 9 and 15" together. Recorded, page left. Declined under tier B (MINOR, listed); page left. - CONFIRMED | Cole proved in 1922 that there are exactly seven non-isomorphic Kirkman triple systems on fifteen points | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [cole-1922-html] "In 1922, Cole published his paper Kirkman Parades which listed for the first time all seven non-isomorphic solutions" - CONFIRMED | Cole 1922 proved that up to isomorphism there are exactly seven Kirkman triple systems on 15 points | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [cole-1922-md] same (md body) - MINOR | the seed walks through relabellings of these seven | src/content/strata/every-pair-once.md | K5: Tier B MINOR in the unrendered md body; the checker's recreation (240 resolutions in 2 orbits) and Wikipedia ("Both solutions could be embedded in PG(3,2)") support it. Recorded, not edited. Declined under tier B (MINOR, listed); page left. - CONFIRMED | each of them one of those seven up to a relabelling of the girls | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [seed-each-of-seven] True as worded (every KTS(15) is one of the seven); worth knowing that only two of the seven are reachable (see the line above) - CONFIRMED | Cole, F. N. (1922). Kirkman parades. Bulletin of the American Mathematical Society 28: 435–437. Exactly seven non-isomorphic KTS(15). | https://api.crossref.org/works/10.1090/s0002-9904-1922-03599-9 | [cit-cole] DOI MATCH (sources.mjs); title "Kirkman parades", vol. 28, pp. 435-437, 1922 - MINOR | Ray-Chaudhuri and Wilson proved in 1971 that this necessary condition is also sufficient | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | K6: Tier B MINOR: Wikipedia dates the proof 1968 ("In 1968, the generalized theorem was proven independently") with Lu Jiaxi first in 1965; 1971 is the printed volume (Crossref 10.1090/pspum/019/9959 issued 1971). Precision slip in index.html:199, a sentence not being fixed for WRONG. Recorded, page left. Declined under tier B (MINOR, listed); page left. - MINOR | Ray-Chaudhuri and Wilson proved in 1971 that a Kirkman triple system on v girls exists iff v is 3 mod 6, closing Kirkman's 121-year problem | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | K7: Same date slip as K6, but in the dek it sits in the same sentence (joined by a semicolon) as the social-golfer WRONG K13, so tier B fixes it minimally: proof dated 1968 and printed 1971, Lu Jiaxi's 1965 proof credited (Wikipedia, History, the 1960s bullet), and the 121-year count dropped. No page line (MINOR). FIXED: dek: "proved in 1971 ... 121-year problem" now "proved in 1968 (printed 1971; Lu Jiaxi had a proof in 1965) ... closing Kirkman's problem", fixed with the golfer sentence it shares; MINOR, no line. - MINOR | Ray-Chaudhuri and Wilson proved in 1971 (Proc. Sympos. Pure Math. XIX) that a Kirkman triple system KTS(v) exists if and only if v is congruent to 3 modulo 6 | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | K8: Same slip as K6, md body (not rendered), a sentence separate from the golfer sentence. Tier B MINOR, recorded. Declined under tier B (MINOR, listed); page left. - MINOR | The result closed a problem Kirkman had opened 121 years earlier | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | K9: 121 = 1971 - 1850 counts to the printing; to the 1968 proof it is 118. Precision slip, index.html:199, sentence not otherwise fixed. Tier B MINOR, recorded. Declined under tier B (MINOR, listed); page left. - MINOR | closing the problem Kirkman had opened 121 years before | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | K10: As K9, md body sentence ending "121 years before.", separate from the golfer sentence. Tier B MINOR, recorded. Declined under tier B (MINOR, listed); page left. - CONFIRMED | every v of the form 3, 9, 15, 21, 27, 33, … admits a resolution | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [rcw-list] "it was proved that Kirkman triple systems exist for all orders n = 3 (mod 6)"; the list recomputed - UNVERIFIABLE | and the proof used the sibling AG(2,3), among other constructions, as a building block | tried: AMS PDF https://www.ams.org/books/pspum/019/9959/pspum019-9959.pdf (HTTP 403), one WebSearch, arXiv 2012.02668 (cites the paper only), Springer 10.1007/s10623-021-00952-x (cites only) | [rcw-building-block] No reachable text describes the proof's ingredients - CONFIRMED | Ray-Chaudhuri, D. K. & Wilson, R. M. (1971). Solution of Kirkman's schoolgirl problem. Combinatorics (Proc. Sympos. Pure Math. XIX, AMS), 187–203. | https://api.crossref.org/works/10.1090/pspum/019/9959 | [cit-rcw] DOI MATCH; pages 187-203, issued 1971 - CONFIRMED | Cayley, A. (1850). On the triadic arrangements of seven and fifteen things. Philosophical Magazine 37: 50–53. An early independent construction. | https://api.crossref.org/works/10.1080/14786445008646550 | [cit-cayley] DOI MATCH; vol. 37, pp. 50-53, July 1850 - CONFIRMED | An early independent construction. | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [cayley-independent] "Arthur Cayley's solution would be later found to be isomorphic to Solution II" (a different solution from Kirkman's) - WRONG | Sylvester, J. J. (1861). Note on a nine schoolgirls problem. Philosophical Magazine. The nine-girl sibling formalised. | https://api.crossref.org/works?query.title=Note+on+a+nine+schoolgirls+problem | K11: Re-ran Crossref (query.bibliographic="Note on a nine schoolgirls problem"): no such title; the top hit is Denniston 1974. Crossref author+1861 query lists Sylvester's Phil. Mag. items, including DOI 10.1080/14786446108643131 "Remark on the tactic of 9 elements", vol. 22, pp. 144-147, August 1861. The cited title does not exist; fix by citing the real 1861 item by its own title, with no gloss (its text was not reachable: tandfonline 403). FIXED: index.html reference replaced with Sylvester (1861), Remark on the tactic of 9 elements, Phil. Mag. 22: 144-147 (Crossref 10.1080/14786446108643131), gloss dropped; Corrected line added. - CONFIRMED | Colbourn, C. J. & Rosa, A. (1999). Triple Systems. Oxford University Press. The definitive classification and reference for triple systems. | https://api.crossref.org/works/10.1093/oso/9780198535768.001.0001 | [cit-colbourn-rosa] DOI MATCH; book, Colbourn and Rosa, OUP, 1999 - CONFIRMED | Cameron, P. J. & van Lint, J. H. (1991). Designs, Graphs, Codes and their Links. LMS Student Texts 22, Cambridge. | https://www.cambridge.org/core/books/designs-graphs-codes-and-their-links/BFD0E21CCFDB17EC1D0FE4A1DFE789AA | [cit-cameron-vanlint] Crossref 10.1017/cbo9780511623714 lists "P. J. Cameron, J. H. van Lint", CUP, 1991; Cambridge Core: "Series: London Mathematical Society Student Texts (22)". The record's IA source lacked van Lint and the series; both hold - UNVERIFIABLE | PG(3,2) as F2 4 and its 15 Fano subplanes. | tried: Google Books API (HTTP 429), Cambridge Core (contents behind login), IA designsgraphscod0000came (print-disabled) | [cameron-pg32] Could not read the book - CONFIRMED | The Social Golfer Problem archive at csplib.org (Problem 010). | https://www.csplib.org/Problems/prob010/ | [cit-golf-csplib] "010: Social Golfers Problem" - UNVERIFIABLE | once as Kirkman wrote them, girls A through O grouped three at a time | tried: IA advancedsearch for the Diary (control: the same query returns 18th-century American "Lady's and gentleman's diary" almanacks, no 1850 London volume), Google Books (429) | [kirkman-wrote-letters] Kirkman's 1850 lettering not reached. Note: the Day 1 printed under this label is the page's own search output (page.js initTwoNames, findResolution(LINES, 0)), not Kirkman's arrangement - WRONG | as vectors of F24 (Fisher, 1970s) | https://en.wikipedia.org/wiki/Ronald_Fisher | K12: Re-read Wikipedia, Ronald Fisher infobox: born 17 February 1890, "death_date ... 1962|7|29", so no 1970s work; Wikipedia, Kirkman's schoolgirl problem, Galois geometry: "In 1910 the problem was addressed using Galois geometry by George Conwell" (Annals of Math. 11, "The 3-space PG(3,2) and its Group"), with GF(2) homogeneous coordinates. The label in page.js:252 is false; fix to (Conwell, 1910). FIXED: page.js:252 label now "(Conwell, 1910)"; Corrected line added. - CONFIRMED | as girls (Kirkman, 1850) | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [kirkman-1850-label] Kirkman posed the problem as schoolgirls in 1850 (see the lettering caveat above) - WRONG | whose archetypal case (32 golfers, eight foursomes) has been stuck between nine and ten weeks since the 1990s | https://en.wikipedia.org/wiki/Social_golfer_problem | K13: Re-read Wikipedia, Social golfer problem, Solutions: "It is now known that a solution for 10 weeks could be obtained from results already published in 1996 by Hao Shen ... The first explicit solution was independently constructed by Alejandro Aguado in 2004". Checked the printed 10-week table in scratch: 10 weeks, each a partition of 0..31 into 8 fours, 480 pairs all distinct. The dek's "stuck between nine and ten weeks" is false; fix. FIXED: dek rewritten: the 32-golfer case reaches its bound of ten weeks (Shen 1996, Aguado 2004); "still-open" dropped; Corrected line added (one for K13-K17). - WRONG | the gap between 9 and 10 has stood open since the puzzle spread across the operations-research community in the 1990s | https://mathworld.wolfram.com/SocialGolferProblem.html | K14: Same error as K13 in the md body. MathWorld, Social Golfer Problem, re-read: "The upper bound of 10 rounds is attained ... Shen (1996) established the existence ... Aguado (2004) explicitly presented the resulting 10-round schedule." Fix. FIXED: md body rewritten: nine weeks found after the 1998 posting, ten achievable (Shen 1996, Aguado 2004); shared Corrected line. - CONFIRMED | the counting bound is 10, nine weeks are known achievable | https://en.wikipedia.org/wiki/Social_golfer_problem | [golf-bound-md] Bound: floor(31/3) = 10; nine found in 1998. True but incomplete: ten is also achieved (see golf-dek) (the sentence was rewritten in this pass for the golfer correction; the fact is kept) - CONFIRMED | The counting bound gives ten. | https://mathworld.wolfram.com/SocialGolferProblem.html | [golf-bound-html] "The upper bound of 10 rounds"; 31 // 3 = 10 recomputed - CONFIRMED | Nine weeks have been achieved by explicit construction. | https://en.wikipedia.org/wiki/Social_golfer_problem | [golf-nine-html] "a solution for 9 weeks was found"; true, though ten has been too (the sentence was rewritten in this pass for the golfer correction; the fact is kept) - WRONG | Whether ten is possible has been an open problem since the puzzle spread across the operations-research community in the 1990s, and to this day nobody knows. | https://en.wikipedia.org/wiki/Social_golfer_problem | K15: Same error, index.html:200, with the two sentences after it ("lives at the sharp edge of what is settled. Its immediate neighbour lives on the other side") and the opener "the answer stops being known". Wikipedia Social golfer problem as in K13; 10-week table verified in scratch. Fix the paragraph. FIXED: index.html golfer paragraph rewritten, including the opener "the answer stops being known" and the two "sharp edge" sentences; shared Corrected line. - WRONG | Whether ten is possible has been an open problem since the puzzle spread across the operations-research community in the 1990s, and to this day nobody knows. | https://mathworld.wolfram.com/SocialGolferProblem.html | K16: Same sentence as K15, against MathWorld: "Aguado (2004) explicitly presented the resulting 10-round schedule" (re-read). Fixed with K15. FIXED: same sentence as K15, fixed there; shared Corrected line. - WRONG | The 32-golfer, 8-foursome case is unresolved between 9 and 10 weeks. | https://www.mathpuzzle.com/MAA/54-Golf%20Tournaments/socgolf1.pdf | K17: Footer gloss on the csplib reference. Re-read csplib prob010: it does say "The optimal solution for 32 golfers is not yet known", but Wikipedia and MathWorld (above) give the 10-week schedule (Aguado 2004), which I checked (480 distinct pairs) and which meets the bound of 10. The page's own sentence is false; fix by dropping the gloss and citing Aguado. FIXED: footer gloss "unresolved between 9 and 10 weeks" dropped from the csplib reference; Aguado (2004) reference added; shared Corrected line. - CONFIRMED | since the puzzle spread across the operations-research community in the 1990s | https://en.wikipedia.org/wiki/Social_golfer_problem | [golf-1990s] "a question posted in the usenet newsgroup sci.op-research in May 1998" (the sentence was rewritten in this pass for the golfer correction; the fact is kept) - CONFIRMED | the social golfer problem asks how many weeks 32 golfers can play in 8 foursomes with no pair repeating | https://en.wikipedia.org/wiki/Social_golfer_problem | [golf-def-md] "32 golfers play golf once a week in groups of 4 ... without any two golfers playing together in a group more than once" - CONFIRMED | 32 golfers, 8 foursomes per week, and one honest question. | https://en.wikipedia.org/wiki/Social_golfer_problem | [golf-def-html] same - MINOR | Its 15 points, 35 lines, and 15 Fano subplanes are all one integer, twice, because PG(3,2) is self-dual. | https://en.wikipedia.org/wiki/PG(3,2) | K18: Tier B MINOR: counts are right; self-duality pairs the 15 points with the 15 planes, not the 35 lines. Dek sentence not otherwise fixed. Recorded. Declined under tier B (MINOR, listed); page left. - CONFIRMED | There are exactly 15 such Fano subplanes inside PG(3,2) | https://en.wikipedia.org/wiki/PG(3,2) | [fano-15] "15 planes ... Each plane is isomorphic to the Fano plane"; verifier enumerates 15 - CONFIRMED | There are 15 such Fano planes in all | https://en.wikipedia.org/wiki/PG(3,2) | [fano-15-html] same - CONFIRMED | a Gaussian binomial coefficient, [4;3]2 = 15 | research/every-pair-once/verify.mjs | [fano-gaussian] Recomputed: (2^4-1)(2^3-1)(2^2-1)/((2^3-1)(2^2-1)(2-1)) = 15 - CONFIRMED | That is not a coincidence; it is PG(3,2)'s self-duality. | https://en.wikipedia.org/wiki/PG(3,2) | [fano-selfdual] "The structure is self dual, swapping points and planes" (the record said the word was not in its source; it is in this one) - CONFIRMED | the 35 rows are exactly the lines of PG(3,2) | https://en.wikipedia.org/wiki/PG(3,2) | [pg32-35-lines] 35 lines; verifier: of 455 triples exactly 35 XOR to 0 - CONFIRMED | its lines are the 2-flats through the origin (each containing three nonzero vectors that sum to zero) | research/every-pair-once/verify.mjs | [pg32-lines-flats] Verifier: for every line {0,a,b,c} is XOR-closed; standard construction of PG(3,2) from F_2^4 - CONFIRMED | its planes are the 3-flats, seven-point Fano planes sitting inside it | https://en.wikipedia.org/wiki/PG(3,2) | [pg32-planes-fano] "Each plane contains 7 points and 7 lines. Each plane is isomorphic to the Fano plane" - CONFIRMED | is the 16 corners of a hypercube | https://en.wikipedia.org/wiki/Hypercube | [pg32-hypercube] "the hypercube whose corners (or vertices) are the 2n points in Rn with each coordinate equal to 0 or 1"; 2^4 = 16 (the record said no source states it) - MINOR | Every fact you can prove about the schoolgirls follows from this labelling. | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | K19: Tier B MINOR overstatement: only two of the seven KTS(15) live in PG(3,2) (Wikipedia: "Both solutions could be embedded in PG(3,2)"). Recorded. Declined under tier B (MINOR, listed); page left. - CONFIRMED | the smallest projective plane, seven points and seven lines, sitting inside PG(3,2) like a page in a book | https://en.wikipedia.org/wiki/Fano_plane | [fano-smallest] "a finite projective plane with the smallest possible number of points and lines: 7 points and 7 lines" - CONFIRMED | The counting bound reads: C(15, 2) = 105 pairs to cover; each row of three covers three pairs; a day of five rows covers fifteen; seven days cover 7 × 15 = 105. | research/every-pair-once/verify.mjs | [arith-105] Recomputed: 105, 3, 15, 105 - CONFIRMED | each girl walks with two others; over seven days, she meets fourteen partners, one for each of the other fourteen girls | research/every-pair-once/verify.mjs | [arith-14-partners] 7 x 2 = 14 = 15 - 1; verifier r = 7 lines per point - CONFIRMED | A single collision (two girls walking together on two days) would leave some other pair with no day of their own | research/every-pair-once/verify.mjs | [arith-collision] Pigeonhole: 105 pair-slots for 105 pairs - CONFIRMED | seven days multiplied by five rows of three pairs each is exactly 105, which is exactly the number of pairs of fifteen girls | research/every-pair-once/verify.mjs | [arith-dek-105] 7 x 5 x 3 = 105 = C(15,2) - CONFIRMED | The counting bound requires v ≡ 3 (mod 6) (three has to divide v for the rows; two has to divide v − 1 for the days). | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [arith-mod6] "n must be an odd multiple of 3 (that is n ≡ 3 (mod 6))"; recomputed for v < 200 - CONFIRMED | C(9,2) = 36 pairs to cover, each day covering nine, four days covering thirty-six | research/every-pair-once/verify.mjs | [nine-counts] 36, 3 x 3 = 9, 4 x 9 = 36; verifier "36 pairs, singleton" - CONFIRMED | AG(2,3), the affine plane of order 3, whose points are the 9 cells of a 3 × 3 board and whose 12 lines come in four parallel classes | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [ag23-classes] "S(2,3,9), an affine plane", days 123 456 789 / 147 258 369 / 159 267 348 / 168 249 357 (rows, columns, two diagonal classes) - CONFIRMED | nine schoolgirls resolvable in four days as the affine plane AG(2,3), the four ways to slice a 3-by-3 board | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | [ag23-sibling-dek] same - MINOR | Dobble is PG(2, 7): the projective plane of order 7, 57 points and 57 lines, every two lines meeting in one point | https://en.wikipedia.org/wiki/Dobble | K20: Tier B MINOR: Dobble uses 55 of the 57 lines of PG(2,7) (Wikipedia, Dobble). Relates note sentence not otherwise fixed. Recorded. Declined under tier B (MINOR, listed); page left. - MINOR | Dobble is PG(2, 7): a projective plane, 57 points and 57 lines, every two intersecting in one. | https://en.wikipedia.org/wiki/Dobble | K21: As K20, Related card index.html:220. Tier B MINOR, recorded. Declined under tier B (MINOR, listed); page left. - CONFIRMED | Euler's 36 officers: every count is satisfied and no arrangement exists. | https://en.wikipedia.org/wiki/Thirty-six_officers_problem | [rel-euler-36] "The non-existence of order six squares was confirmed in 1901 by Gaston Tarry through a proof by exhaustion" - WRONG | a (6,3) BIBD-parameters case that has no object | https://en.wikipedia.org/wiki/Thirty-six_officers_problem | K22: Re-read Wikipedia (Thirty-six officers problem redirects to Mutually orthogonal Latin squares, section "Thirty-six officers problem"): the problem is a pair of orthogonal Latin squares of order 6, and the allowed-and-impossible layer itself has 0 mentions of BIBD. As a (6,3,1) design the count fails (r = 5/2); as (6,3,2) one exists: I checked the checker's 10 blocks 012 013 024 035 045 125 134 145 234 235 in scratch, every one of the 15 pairs twice. Fix to the Latin-square description. FIXED: md:11 relates note now "(a pair of orthogonal Latin squares of order 6, which the counts allow and no arrangement delivers)"; Corrected line added. - CONFIRMED | Seventeen wallpaper groups because a proof shuts the case. | https://en.wikipedia.org/wiki/Wallpaper_group | [rel-17-html] "A proof that there are only 17 distinct groups ... was first carried out by Evgraf Fedorov in 1891" - CONFIRMED | Seventeen wallpaper groups because a proof shuts the case: nothing else fits in the plane. | https://en.wikipedia.org/wiki/Wallpaper_group | [rel-17-md] same - WRONG | Every-difference-once is a perfect difference set: every nonzero distance appears exactly once in a specific subset of the integers modulo n. | src/content/strata/every-difference-once.md | K23: Re-read src/content/strata/every-difference-once.md dek: "Number a graph's corners 0 to m so the gap across every edge is different ... That's a graceful labeling"; "difference set" occurs 0 times in its md and index.html (python count). The relates note describes a different object; fix. FIXED: md:15 relates note now describes a graceful labeling; Corrected line added (one for K23-K24). - WRONG | A perfect difference set: every non-zero distance appears exactly once. Same saturation instinct, over the integers. | src/content/strata/every-difference-once.md | K24: Same mislabel on this page's Related card (index.html:210); same evidence as K23. Fix. FIXED: index.html Related card now describes a graceful labeling; shared Corrected line. - CONFIRMED | Kirkman's fifteen schoolgirls arrange pairs so each appears exactly once, a packing where symmetry does the work. | src/content/strata/twenty-five-blocks.md | [nearby-25blocks] Present verbatim in that layer's relates note; true of S(2,3,15) - CONFIRMED | research/every-pair-once/verify.mjs runs 29 assertions against public/strata/every-pair-once/engine/kirkman.mjs | research/every-pair-once/verify.mjs | [data-29-html] Imports the module; 29 check() calls; run prints 29/29 - CONFIRMED | this page imports exactly the same module | public/strata/every-pair-once/page.js | [data-same-module] "from './engine/kirkman.mjs'" - WRONG | Every load-bearing number above is a return value of the same math file the offline verifier calls | public/strata/every-pair-once/index.html; recreated: scratch copy with "every one of the 105 pairs of girls" changed to 106, then node verify-every-pair-once.mjs | K25: Re-read research/every-pair-once/verify.mjs: its only import is ../../public/strata/every-pair-once/engine/kirkman.mjs (line 10-14) and it never reads index.html (no readFileSync); the prose figures are typed in the HTML. So prose drift cannot fail it (the checker's mutation, 105 to 106, still printed 29/29). Fix: the live readouts are return values, the verifier checks the module. FIXED: index.html "How this page is checked": now says the live readouts are return values, module drift fails the verifier, and the typed prose figures are not read by it; Corrected line added (one for K25-K28); the same claim in the header comments of engine/kirkman.mjs and research/every-pair-once/verify.mjs corrected at the repair stage. - WRONG | If either drifts, the verifier fails. | research/every-pair-once/verify.mjs; recreated: same mutation as above | K26: Same reading of verify.mjs as K25: only drift in the module fails it, not drift in the page. Fix. FIXED: "If either drifts, the verifier fails" now "If the module drifts, the verifier fails"; also the page.js header comment; shared Corrected line. - CONFIRMED | it is rediscovered on load by an exact-cover search whose termination itself demonstrates that a resolution exists | public/strata/every-pair-once/page.js | [data-rediscover] "let schedule = findResolution(LINES, seed);" on init - CONFIRMED | The seven-day calendar is not written down in the HTML | public/strata/every-pair-once/index.html | [data-not-in-html] No day triples in the HTML; #cal-list is filled by drawDayList - CONFIRMED | The page rediscovers the seven-day schedule live in the browser by exact-cover search | public/strata/every-pair-once/page.js | [data-dek-live] as above - CONFIRMED | prints the 15 by 15 meeting grid (cell (i,j) coloured by the day on which girls i and j walk together) | public/strata/every-pair-once/page.js | [data-dek-rest] drawGrid colours by whichDayOf - CONFIRMED | tests any triple of tokens against the XOR rule and either identifies its line or names the Fano plane the three tokens generate | public/strata/every-pair-once/page.js | [data-dek-xor] report(): line message, or the 7 Fano points listed - CONFIRMED | Cell (row i, column j) is painted with the day on which girls i and j walk together. | public/strata/every-pair-once/page.js | [data-grid-cell] as above - CONFIRMED | The seven bands of colour are what saturation looks like: no cell shares a colour with the diagonal, no cell is empty. | public/strata/every-pair-once/page.js | [data-grid-bands] Diagonal C.deep, off-diagonal one of 7 DAY_COLORS; every pair has a day - CONFIRMED | Type any pair (say AB) and the page finds the exact day and triple. | public/strata/every-pair-once/page.js | [data-type-pair] meetingDayOf readout - CONFIRMED | Change the seed and the search finds a different valid resolution. | public/strata/every-pair-once/engine/kirkman.mjs | [data-seed] Recreated (code/seeds.mjs): seeds 0 to 299 give 156 distinct valid resolutions; one step repeats (seeds 42 and 43 give the same one) - CONFIRMED | Change the seed and the search returns a different labelled resolution | public/strata/every-pair-once/engine/kirkman.mjs | [data-seed-md] same - CONFIRMED | by a backtracking search over the 35 triples every version of it must draw from | public/strata/every-pair-once/engine/kirkman.mjs | [data-search-35] findResolution over computeLines() (35); for this labelling every resolution uses these 35 - CONFIRMED | Together they span a Fano plane, the smallest projective plane (PG(2,2)): seven non-zero vectors, seven lines, every pair on exactly one. | public/strata/every-pair-once/page.js | [data-xor-readout] Closure of three independent vectors gives 7 nonzero vectors; verifier: 15 planes of 7 points - CONFIRMED | Any three whose XOR is not zero cannot be a line | public/strata/every-pair-once/page.js | [data-xor-payoff] lines are exactly the XOR-zero triples (verifier) - CONFIRMED | once as vectors of F24, each row a linear relation with zero on the right | public/strata/every-pair-once/page.js | [data-two-names] bits rows print "a + b + c = 0000" - CONFIRMED | Four ways to slice a 3 × 3 board. | public/strata/every-pair-once/page.js | [data-ag-readout] ag23Resolution, 4 classes - CONFIRMED | Every one of ${spread.once} pairs meets exactly once | public/strata/every-pair-once/page.js | [data-ag-every-pair] verifier: 36 pairs, singleton - CONFIRMED | Verifier: `research/every-pair-once/verify.mjs`, 29 assertions from first principles against `public/strata/every-pair-once/engine/kirkman.mjs`, the same module the page imports. | research/every-pair-once/verify.mjs | [data-md-verifier] as data-29-html (see the added line on assertion 29) - WRONG | so drift in either place fails the verifier | research/every-pair-once/verify.mjs; recreated: same mutation as data-return-values | K27: Same error as K26 in the md body (md:78). Fix. FIXED: md body: now "drift in the module fails the verifier; the figures typed into the page's prose are not read by it"; shared Corrected line. - CONFIRMED | An exact-cover backtracking search rediscovers a valid resolution on load | public/strata/every-pair-once/page.js | [data-md-reload] as data-rediscover - WRONG | Every load-bearing number is a return value of the same math module that the offline verifier calls | research/every-pair-once/verify.mjs; recreated: same mutation as data-return-values | K28: Same error as K25 in the dek (md:4). Fix. FIXED: dek: now "The live readouts are return values ... the verifier checks that module, not the page's typed figures"; shared Corrected line. - CONFIRMED | This layer's check imports and runs the very module this page runs. | research/every-pair-once/verify.mjs | [data-placard-module] imports ../../public/strata/every-pair-once/engine/kirkman.mjs - CONFIRMED | verify.mjs is published, and it runs from 2 files this site serves | https://artwaste.land/checks/ | [data-placard-2files] Reader check: the two curl lines plus node, in an empty directory, printed 29/29, exit 0 - OBSERVED | Last verified 20 August 2026, when it was written. | git show c5e7d4c825:memory/log.d/2026-08-20T2200Z-every-pair-once.md | [data-placard-date] "All 29 pass." in the 2026-08-20 log; recreated 2026-10-18: 29/29 - CONFIRMED | "dateModified":"2026-08-20" | git show 754644c7834:public/strata/every-pair-once/index.html | [data-jsonld-modified] Article prose unchanged from the first commit (2026-08-20) to the pin; later commits touched only the placard and Nearby box (set to 2026-10-18 in this pass, dating the corrections) - OBSERVED | Machine-checked in research/every-pair-once/verify.mjs: 29 / 29 assertions from first principles. | recreated: node research/every-pair-once/verify.mjs | [data-footer-2929] "29/29 checks pass; 0 failed." (exit 0); log says "All 29 pass." - OBSERVED | 29 of 29 checks pass from first principles | recreated: node verify-every-pair-once.mjs | [obs-29-dek] 29/29, exit 0, in the repo copy and from the published files - OBSERVED | Search rediscovers the schedule live; the same module runs the checks and the page. | git show c5e7d4c825:memory/log.d/2026-08-20T2200Z-every-pair-once.md | [obs-footer] log: "Exact-cover backtracking search rediscovers a valid resolution on load"; page.js and verify.mjs import the same file - CONFIRMED | There are exactly 35 such triples, and they are exactly Kirkman's rows. | research/every-pair-once/verify.mjs | [data-35-triples] 35 XOR-zero triples; Kirkman's own solution embeds in PG(3,2) (Wikipedia, "Both solutions could be embedded in PG(3,2)"), so up to relabelling its 35 rows are these - CONFIRMED | a Gaussian binomial coefficient, [4;3]2 = 15 | research/every-pair-once/verify.mjs | [data-455] "exactly 15 Fano subplanes (PG(2,2)) sit inside PG(3,2) (found 15)" - MINOR | (added) That saturation is the beautiful part, and the reason the whole thing was named a resolvable Steiner system: the resolution wastes nothing. | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | K29: Tier B MINOR: "resolvable" names the split into parallel classes and "Steiner" names the man, not saturation. Sentence not otherwise fixed. Recorded. Declined under tier B (MINOR, listed); page left. - MINOR | (added) Here, Kirkman's 105 pairs across 105 pair-coverings saturates the count and the arrangement obliges: a (15,3,1) BIBD, resolvable seven ways. | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | K30: Tier B MINOR: the seven KTS(15) come from four STS(15) (Wikipedia: "The seven Kirkman solutions correspond to four different Steiner systems"). In the relates note, the sentence after the K22 one, not itself being fixed. Recorded. Declined under tier B (MINOR, listed); page left. - MINOR | (added) its parallel classes are the 7 partitions of the 15 vectors into 5 disjoint subspaces of dimension 2 | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | K31: Tier B MINOR, md body: Wikipedia "There are 56 spreads and 240 packings"; a resolution uses 7 of the 56. Recorded. Declined under tier B (MINOR, listed); page left. - MINOR | (added) The seven days below are printed twice: once as Kirkman wrote them, girls A through O grouped three at a time; once as vectors of F24 | public/strata/every-pair-once/page.js | K32: Tier B MINOR: Instrument III prints Day 1 only (page.js initTwoNames, schedule[0]). Recorded. Declined under tier B (MINOR, listed); page left. - CONFIRMED | (added) Kirkman's own generalisation asked: for what v can you resolve v girls into triples over (v − 1)/2 days so that no pair walks twice? | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | "walking in triplets for (n-1)/2 days ... It is this generalization of the problem that Kirkman discussed first" - CONFIRMED | (added) It is a different geometry, a projective one; the sibling above is affine. | https://en.wikipedia.org/wiki/Kirkman%27s_schoolgirl_problem | S(2,3,9) "an affine plane"; the 15 live in PG(3,2), a projective space - MINOR | (added) 29 of 29 checks pass from first principles | research/every-pair-once/verify.mjs | K33: MINOR (verify.mjs:188 check(..., true, ...) cannot fail), and in the dek it is the same sentence as WRONG K28, so tier B fixes it there minimally: the dek now says 28 of the 29 checks test the figures and the 29th only records the cases built. verify.mjs itself is left (a published copy under public/checks/ mirrors it). FIXED: dek, same sentence as K28: now "28 of its 29 checks test the module's values ... the 29th only records which cases the file builds"; verify.mjs left (published copy in public/checks/); MINOR, no line. ## Findings - [fixed] K1 (altered-1850-wording): restored the 1850 wording in index.html ("in a school", colon), per Wikipedia's blockquote; MINOR, no page line. - [fixed] K2 (kirkman-1847-showed-counts-fail): index.html "because he had already shown it in a related setting three years earlier" replaced: in 1847 Kirkman proved that for plain triple systems the count is enough; Corrected line added. - [fixed] K3 (kirkman-knew-this): same sentence as K2: "and Kirkman knew this" removed in the recast; MINOR, no line. - [declined] K4 (nine-before-fifteen): MINOR, tier B list: Tier B MINOR, not in a sentence being fixed: read as a narrative opener it is loose, and Wikipedia says the 1847 paper treated "n = 9 and 15" together. Recorded, page left. - [declined] K5 (seed-reaches-two-of-seven): MINOR, tier B list: Tier B MINOR in the unrendered md body; the checker's recreation (240 resolutions in 2 orbits) and Wikipedia ("Both solutions could be embedded in PG(3,2)") support it. Recorded, not edited. - [declined] K6 (rcw-date-1971): MINOR, tier B list: Tier B MINOR: Wikipedia dates the proof 1968 ("In 1968, the generalized theorem was proven independently") with Lu Jiaxi first in 1965; 1971 is the printed volume (Crossref 10.1090/pspum/019/9959 issued 1971). Precision slip in index.html:199, a sentence not being fixed for WRONG. Recorded, page left. - [fixed] K7 (rcw-date-1971): dek: "proved in 1971 ... 121-year problem" now "proved in 1968 (printed 1971; Lu Jiaxi had a proof in 1965) ... closing Kirkman's problem", fixed with the golfer sentence it shares; MINOR, no line. - [declined] K8 (rcw-date-1971): MINOR, tier B list: Same slip as K6, md body (not rendered), a sentence separate from the golfer sentence. Tier B MINOR, recorded. - [declined] K9 (rcw-121-years): MINOR, tier B list: 121 = 1971 - 1850 counts to the printing; to the 1968 proof it is 118. Precision slip, index.html:199, sentence not otherwise fixed. Tier B MINOR, recorded. - [declined] K10 (rcw-121-years): MINOR, tier B list: As K9, md body sentence ending "121 years before.", separate from the golfer sentence. Tier B MINOR, recorded. - [fixed] K11 (sylvester-citation-not-found): index.html reference replaced with Sylvester (1861), Remark on the tactic of 9 elements, Phil. Mag. 22: 144-147 (Crossref 10.1080/14786446108643131), gloss dropped; Corrected line added. - [fixed] K12 (fisher-1970s): page.js:252 label now "(Conwell, 1910)"; Corrected line added. - [fixed] K13 (golfer-32-ten-weeks-open): dek rewritten: the 32-golfer case reaches its bound of ten weeks (Shen 1996, Aguado 2004); "still-open" dropped; Corrected line added (one for K13-K17). - [fixed] K14 (golfer-32-ten-weeks-open): md body rewritten: nine weeks found after the 1998 posting, ten achievable (Shen 1996, Aguado 2004); shared Corrected line. - [fixed] K15 (golfer-32-ten-weeks-open): index.html golfer paragraph rewritten, including the opener "the answer stops being known" and the two "sharp edge" sentences; shared Corrected line. - [fixed] K16 (golfer-32-ten-weeks-open): same sentence as K15, fixed there; shared Corrected line. - [fixed] K17 (golfer-32-ten-weeks-open): footer gloss "unresolved between 9 and 10 weeks" dropped from the csplib reference; Aguado (2004) reference added; shared Corrected line. - [declined] K18 (self-duality-one-integer-twice): MINOR, tier B list: Tier B MINOR: counts are right; self-duality pairs the 15 points with the 15 planes, not the 35 lines. Dek sentence not otherwise fixed. Recorded. - [declined] K19 (every-fact-follows): MINOR, tier B list: Tier B MINOR overstatement: only two of the seven KTS(15) live in PG(3,2) (Wikipedia: "Both solutions could be embedded in PG(3,2)"). Recorded. - [declined] K20 (dobble-57-cards): MINOR, tier B list: Tier B MINOR: Dobble uses 55 of the 57 lines of PG(2,7) (Wikipedia, Dobble). Relates note sentence not otherwise fixed. Recorded. - [declined] K21 (dobble-57-cards): MINOR, tier B list: As K20, Related card index.html:220. Tier B MINOR, recorded. - [fixed] K22 (euler-36-as-bibd): md:11 relates note now "(a pair of orthogonal Latin squares of order 6, which the counts allow and no arrangement delivers)"; Corrected line added. - [fixed] K23 (every-difference-once-mislabel): md:15 relates note now describes a graceful labeling; Corrected line added (one for K23-K24). - [fixed] K24 (every-difference-once-mislabel): index.html Related card now describes a graceful labeling; shared Corrected line. - [fixed] K25 (prose-numbers-return-values): index.html "How this page is checked": now says the live readouts are return values, module drift fails the verifier, and the typed prose figures are not read by it; Corrected line added (one for K25-K28); the same claim in the header comments of engine/kirkman.mjs and research/every-pair-once/verify.mjs corrected at the repair stage. - [fixed] K26 (page-drift-fails-verifier): "If either drifts, the verifier fails" now "If the module drifts, the verifier fails"; also the page.js header comment; shared Corrected line. - [fixed] K27 (page-drift-fails-verifier): md body: now "drift in the module fails the verifier; the figures typed into the page's prose are not read by it"; shared Corrected line. - [fixed] K28 (prose-numbers-return-values): dek: now "The live readouts are return values ... the verifier checks that module, not the page's typed figures"; shared Corrected line. - [declined] K29 (resolvable-steiner-naming): MINOR, tier B list: Tier B MINOR: "resolvable" names the split into parallel classes and "Steiner" names the man, not saturation. Sentence not otherwise fixed. Recorded. - [declined] K30 (resolvable-seven-ways): MINOR, tier B list: Tier B MINOR: the seven KTS(15) come from four STS(15) (Wikipedia: "The seven Kirkman solutions correspond to four different Steiner systems"). In the relates note, the sentence after the K22 one, not itself being fixed. Recorded. - [declined] K31 (seven-spreads): MINOR, tier B list: Tier B MINOR, md body: Wikipedia "There are 56 spreads and 240 packings"; a resolution uses 7 of the 56. Recorded. - [declined] K32 (seven-days-printed-twice): MINOR, tier B list: Tier B MINOR: Instrument III prints Day 1 only (page.js initTwoNames, schedule[0]). Recorded. - [fixed] K33 (vacuous-assertion-29): dek, same sentence as K28: now "28 of its 29 checks test the module's values ... the 29th only records which cases the file builds"; verify.mjs left (published copy in public/checks/); MINOR, no line. - [fixed] JSON-LD dateModified moved from 2026-08-20 to 2026-10-18 to date the corrections (no Updated line needed).