ASTRA v0.3.0 - accessible browser worksheet

Seam, observation, and certificate discipline

A true local result may still certify too little. Separate physical effect from causal legibility, then close the global boundary.

Complete any fields that help the audit. Nothing is required. Use your browser's Print command to make a local paper or PDF copy. The print layout removes navigation and preserves native field values.

Download the compact one-page print PDF. It is static, not electronically fillable, and is not claimed as PDF/UA-conformant.

1. Atomic claim and typed chain

State one claim at one scale. Name the edge types instead of allowing physical, control, observation, archival, or verification roles to substitute for one another.

One sentence with scale, location, date, and one canonical evidence label.
Source, quantity, constitutive law, coupling, sink, and dissipation.
Sensor or controller, target, timing, cost, and failure mode.
Instrument, geometry, sampling path, preprocessing, noise, and missingness.
Deposition, alteration, exposure, recovery, and recognition.
Exact bounded conclusion, assumptions, and detectable failures.
2. Dual rent
Name the studied seam and the reference seam or protocol.
Organism, instrument, archive, analyst, algorithm, proof system, or other declared channel.
What future changes under intervention on the seam? Which outcome distance measures it?
Which candidate generators become more or less distinguishable, and with what gain sign?
Which intervention or new protocol should reverse the masking or diagnostic gain?
Which operation could magnify the diagnostic distinction?
3. Five FOG coordinates and equifinality
Relative to which material, mode, population, or baseline?
Which scales, frequencies, classes, and uncertainties are averaged out?
Which ordinary and extraordinary histories fit the datum?
Is causal order resolved at the mechanism's timescale?
Which costs, reservoirs, denominators, or escape paths are excluded?
Which different histories reach the same physical state?
Which different states become the same measured record?
4. Local-to-global certificate stack

A local certificate may be correct while the global claim remains unsupported. Record each certificate separately.

Differential, assay, mode, morphology, congruence, or bounded fit.
Complete preimages, viable alternatives, causal chain, and residue graph.
Properness, system closure, omitted reservoirs, and asymptotic paths.
Good and bad primes, valuations, extensions, lifting, and identifiers.
Proof object, exact identity, deterministic test, hashes, and scope.
A channel not physically or institutionally downstream of the first.
5. Residue, decision, and failure memory
Location, medium, intensity, covariance, and survival or detection probability.
Specify the observation before the search.
Which result would materially lower confidence?
Prediction ID, test, result, revision, and assumptions added.

Evidence label

Dual rent paid

Global closure adequate

Protocol preregistered

Guardrails

Read the dual-rent and local-to-global explanation, or return to the v0.3.0 release landing page. The internal verification report documents release checks but is not peer review or outside endorsement.