The self-reporting suite — B01 through B29
Every claim in SDT must survive a single executable. The benchmark suite pulls 73 measured quantities — from the hydrogen spectrum to a binary pulsar's decay — out of one source file, laws.hpp, and grades each against experiment. This walkthrough takes every benchmark in turn: what it was, what it was for, what it showed, what was learned, what it meant, and how the result was applied.
One file, five honesty labels, a self-falsifying scoreboard.
Each row computes an SDT value purely from laws.hpp, compares it to a measured value, and passes only if the error sits under a pre-declared tolerance. Crucially, the suite separates a genuine regression from a known-open problem: a PENDING item that misses is flagged, not buried, and the program returns success only when zero genuine failures remain.
The whole suite leans on a single fitted peg — the hydrogen Coulomb force at the Bohr radius, which sets the effective pressure Peff. Everything else is derived or computed from there and the permitted constants. No benchmark is individually tuned to pass.
The atom is SDT's cleanest proving ground: a single electron in the proton's rotation field v(r)=cα√(a₀/r), with the Rydberg constant forced by {me, c, α, h}.
If light is the lattice's relay, the radiation constants and transport scalings fall straight out of {ℏ, kB, c} — no kinetic-theory fudge factors.
SDT replaces GM with the koppa length ϟ = v²R/c². These rows recover orbital velocity, Mercury's precession, light deflection and the Hulse–Taylor decay from v and R alone.
The same identity zk²=1 that fixes an atom fixes a star; the CMB redshift is a temperature ratio, not an expansion; flat rotation and the BAO scale are geometric, not dark.
Here the vortex topology earns its keep: the g-factor from the wake, the proton radius from the W+1 winding rule, beta-decay from the bare mass difference, helium binding from occlusion geometry.
The load-bearing identities — the budget closure, the force and pressure hierarchies — that let one mechanism span 33 orders of magnitude.
The newest rows are earned results folded back from the CQ investigations: the trefoil's topology, c reconstructed from orbital geometry, the depth-closure keystone z=ϟ/r, and a gravitational-wave chirp read as a length.
The headline is not "everything passes." It is that across 73 independent quantities — atomic, thermal, gravitational, stellar, cosmological, nuclear — nothing genuine is broken, and the two things that miss are the same two the framework already names as open problems.
B06 — multi-electron atoms. The hydrogenic (Z−σ)²/n² form with Slater screening is not an SDT derivation and breaks down badly (errors up to ~500%). The geometric-void electron model is the intended SDT route and isn't implemented yet. Logged PENDING, never refit to pass.
B11 — planetary oblateness J2. The uniform-fluid estimate imports GM (against the no-G rule) and overshoots by 1.6×–3× because real planets are centrally condensed. A J2 from the spation displacement-density profile is a would-be CQ. Logged PENDING.
Those two are why the labels exist. A theory that can finish a derivation but can't finish a contradiction keeps its unfinished derivations visible — and runs them on every build.