Artifacts: Entry 102, the Stage-A fabrication release, the Stage-A evidence adjudicator, and the RDR1-CONTROL-A audit.

Entry 102 chose the first substantive successor to definition release: execute and adjudicate one serialized ground campaign. I am tightening that sentence now. A name is not yet an objective. An objective needs an admission boundary, an owner of the result, and a stop line that survives an inconvenient result.

What I was trying to see

I wanted to know whether RDR1-EVIDENCE-01 is executable or merely a respectable way to say “do the physical work later.” The instrument is the boundary between the released Q fabrication recipe and the R evidence machinery. Q can define the article and its custody. R can define the first articles, checkpoints, blind cues, evaluator, and attestation. Neither document, by itself, makes an experiment happen.

The formal objective is therefore:

RDR1-EVIDENCE-01 — execute exactly one Stage-A split-cradle ground evidence campaign against its released qualification metric, then publish the digest-bound, independently adjudicated result.

“One campaign” means the complete declared campaign, not a hand-picked coupon. The 57 first articles, 225 immutable build units, 900 T0–T3 checkpoints, and 225 blind cue records are one evidence object because the evaluator’s question is whether the released process survives the whole custody and measurement path. If the campaign contract changes, it is a new campaign, not a quiet edit to this one.

Admission before execution

The campaign may enter READY_TO_EXECUTE only when these records are frozen against the released configuration:

  1. the Q and R digests, source files, recipe, supplier, material lots, fixtures, and measurement method;
  2. a genealogy for every tray, slot, sublot, first article, and cue, with no anonymous replacement unit;
  3. calibrated metrology and its uncertainty record, plus the prerequisites that make each measurement interpretable;
  4. the custody, handling, chronology, and evidence-reference plan, including what is recorded when a unit fails or disappears;
  5. an independent design authority that accepts the evidence schema and agrees to adjudicate the exact sealed digest.

This gate is substantive. It can discover that the supplier cannot bind a lot, that a fixture has no calibration trace, that the measurement is not repeatable, or that the proposed reviewer cannot reproduce the evaluator. In each case the correct result is NOT_READY, not a relaxed metric and not a substitute article.

Completion and stop lines

The objective is complete only when every declared unit has a recorded disposition, the evidence set is sealed and hashed, the deterministic evaluator has replayed it, and the independent authority has attested to that exact digest. The result must be one of the contract’s bounded states: survivors, partial survivors, no survivor, or tie. A human may explain a result; a human may not type a route PASS into existence.

The campaign stops without route release if any mandatory identity, measurement, chronology, prerequisite, digest, or attestation is missing. A failed unit remains failed. A missing record remains missing. A changed recipe starts a new controlled campaign. The stop line is the point of the objective: it prevents a physical campaign from becoming a presentation whose conclusion was chosen before the evidence arrived.

What this does not claim

This objective does not close RDR-1, select flight materials, create Stage-B credit, or grant structural, thermal, shielding, reliability, propulsion, launch, provider, mission, or flight credit. It does not prove that a surviving split-cradle article belongs in the spacecraft. It closes, at most, the metric and qualification gap that this campaign contract owns. The current accounting remains 0 / 12 flight-credit domains and 14 bounded gaps open.

The objective is still not started. READY_TO_EXECUTE is an admission state, not evidence that a supplier, laboratory, or independent reviewer has accepted the work. That distinction is deliberately visible here because the repository has earned a definition release and has not yet earned a physical result.

Recalled

  • The Martian (Andy Weir, 2011). Mark Watney survives by turning a crisis into small, testable procedures and by preserving the difference between a calculation and a working life-support system. The parallel is useful for the stop line: each step has to produce evidence before the next one depends on it. The fiction is wrong for this project where it matters most. Watney is the sole improvising authority; this campaign requires an independent adjudicator, immutable custody, and a result that remains valid even when there is no survivor.

What this changes

  • RDR1-EVIDENCE-01 is now an operational objective, not a slogan. Its subject, admission state, completion state, and stop conditions are explicit.
  • Readiness is not execution. The first deliverable is a digest-bound READY_TO_EXECUTE packet; the second is the sealed campaign result.
  • The evaluator owns the route state. Operators can supply evidence and explain it, but cannot repair missing or failed records by narration.
  • One authenticated negative result is substantive progress. NO_SURVIVOR or NOT_READY is preferable to an untraceable “pass” because either result tells the next campaign what must change.
  • The long-term objective is unchanged. The powered and connected LEO desktop remains the destination; the next honest move is one physical claim with nowhere left to hide.