nx_boardcontra.nx

nx_boardcontra.nx -- TWO BOARDS, ONE SYMBOL, OPPOSITE CLAIMS (2026-09-04) THE DEFECT, MEASURED. /compare/procgen logged nx_gi_path as "ABSENT: no source exists" at epoch 1788371100. /compare/graphics logged the SAME organ BUILT, PROMOTED and REGISTERED at 1788377700 -- 6,600 seconds later, the same day. Nothing propagated the change back and the first board stood wrong for two days. A seat reading only procgen would have rebuilt a capability that was already live. EVERY EXISTING INSTRUMENT ON THIS SURFACE IS PER-BOARD. The emitter re-measures each watch cell, the contract gate classifies each row, the refs gate checks each citation -- all of them WITHIN one board. A contradiction that only exists BETWEEN two boards is invisible to every one of them, which is why it survived: no instrument was wrong, the question was simply never asked. THE MECHANICAL FORM OF THAT QUESTION: one symbol is declared an OPEN CONTRACT (_ABSENT_:<sym>) on board A while board B carries the SAME symbol as a PRESENT capability. Both cannot be true. This is deliberately narrower than the prose case that motivated it -- it needs no natural language at all -- and it is the half that can be decided from the DATA rather than from sentences. WHY NOT AN AXIS INSIDE nx_board_contract_gate: that organ is proven 12/12 and its two-pass table sizing is tuned to CONTRACT rows only. This question needs every row of every board, present ones included, so it is a different population, not a different predicate over the same one. Extending the incumbent would have meant widening its tables and re-proving its partition; a separate organ leaves that proof intact. CHECK-BEFORE-BUILD RUN AND ADJUDICATED: nx_spendgate's top hit was nx_atlas_contra at the maximum score, READ RATHER THAN ACCEPTED -- it is a two-integer rigor comparator for the knowledge atlas and scores on the word "contra" alone. Not a duplicate. That is the spendgate's own documented behaviour: the top hit is frequently not the answer, which is why it publishes a list and refuses a verdict. PUBLISHES PAIRS, NEVER A WINNER. Which board is right is a judgement about the world; this organ only proves the two disagree. Naming a winner would need a rigor model this has no grounds for. exit: 0 scanned (with or without hits) . 3 no board list readable license_tier: ORIGINAL No hw writes.