# vendor_rows1517_offslice/xy/ -- the two-variable (x, y) connection of the ice-cream cone off the symmetric slice (rows 15/17)

The objects icc-evaluate.py --offslice X Y reads (two levels up in the bundle), every one sha256-pinned there and listed in
XY_MANIFEST.sha256 beside (`sha256sum -c XY_MANIFEST.sha256` in this directory); the seed of record is the served
../line_m3/seed_e5.json (reused by pin, no second copy) and the parser the served ../amf_result.py.

## What is here
- A_xy_eps_monomial.json: the exact two-variable connection (A_x, A_y)(eps, x, y) at w = -1, m^2 = 1, d = 2 - 2 eps (9 x 9, 36 + 36
  nonzero entries, exact rational monomials); integrable (dA = A ^ A vanishes identically in Q(d, x, y): the verdict in A_point.json).
- A_point.json: the export record of that connection: the masters, the denominator census (the factors at d = 2 that the refusals and the
  step control read), the integrability verdict.
- transport_xy.py: the transport engine (straight segments in (x, y), complex waypoints for the detour off x = y); compare_xy.py: the gate.
- S34e5_g60.json: the goal-60 AMFlow twin of the seed at (-3, -4) (--seed-twin); PAe5_g40.json / PAe5_g60.json, PBe5_g40.json /
  PBe5_g60.json, PMe5_g40.json: the independent AMFlow values (eps_order 5, eps^-2..eps^1) at the three reference points.

## What is established (the record's gate; the floors icc-evaluate.py --offslice names and re-runs against the files above, bar 30)
- PA = (-3, -15/4): 48 digits at worst vs the goal-40 value (measured 48.12 d, master [1, 1, 1, 1, -2, 0, 0] at eps^1; 27 pairs);
  vs the goal-60 twin 47 (measured 47.78 d); the eps^1 rows 48, the eps^0 rows 55.
- PB = (-5/2, -15/4): 47 digits at worst vs the goal-40 value (measured 47.94 d, master [1, 1, 1, 1, -2, 0, 0] at eps^1);
  vs the goal-60 twin 47 (measured 47.94 d); the eps^1 rows 47, the eps^0 rows 55.
- PM = (-15/4, -5/2): 47 digits at worst vs the goal-40 value (measured 47.86 d, master [1, -2, 1, 1, 0, 0, 0] at eps^1);
  the eps^1 rows 47, the eps^0 rows 55.
- the two-precision transport pairs (dps 45 / 60): 49 / 49 / 49; the seed twins (goal 40 vs 60 seeds transported): 47 / 47 / 47;
  the AMFlow twins goal 40 vs 60 at PA / PB: 47 / 48.
- Every count is a floor of a two-string agreement; the goal-40 eps_order-5 seed caps the transported values (never a precision claim
  on the AMFlow records).

## What is NOT established
- no goal-60 twin at PM (the mirror point carries one goal-40 value; its two-precision pair is the dps 45/60 transport pair);
- nothing beyond eps^1 (eps_order 5 / target_eps_power 1: eps^-4..eps^1 solved, eps^-2..eps^1 transported and gated);
- no value ON the slice x = y (the nine-master basis degenerates there; icc-evaluate.py's own tiers cover the slice) and no AMFlow-free
  off-slice value (the seed is the AMFlow value of record at (-3, -4)).