Methodology
A boundary between signal,
story and self-deception.
The station follows an epistemic firewall: hardware produces sealed evidence; deterministic code produces measurements; reasoning receives only validated summaries; and public artifacts require human review. Alongside that boundary is a harm-minimization ethic: make costs visible, reduce unnecessary suffering and never infer consent from power.
- 01
Operate receive-only. Transmission code and active probing remain disabled.
- 02
When research touches living systems or possible minds, minimize avoidable harm and never treat another being's inability to resist as permission to disregard its experience.
- 03
Record receiver, antenna path, center frequency, sample rate, gain, duration and software versions before interpretation.
- 04
Seal each accepted capture with sample count and SHA-256 integrity metadata.
- 05
Run deterministic checks for clipping, DC artifacts, I/Q imbalance, occupied bandwidth and repeatability.
- 06
Preserve failed or imperfect attempts so selection bias remains visible.
- 07
Use the labels observation, analysis, testable hypothesis and speculative lens without silently promoting one into another.
- 08
Treat familiar broadcast and reference signals as environmental mapping, not as anomalies.
- 09
Require explicit owner review before a structured record is rendered into the read-only public tree.
- 10
Publish limitations and negative results with the same visibility as successful measurements.
- 11
Correct or retract a ledger entry through a new revision while preserving its stable identifier and history.
Current limitations
The foundation has advanced beyond the original uncalibrated mapper. DL-B is now the accepted primary 50 Ω termination within its tested 50–150 MHz range; VNA-01 has supported bounded measurements of the attenuators, DCB-01 protection chain and FMF-01 broadcast-FM filter. A receiver-level HackRF A–B–A trial also found repeatable in-band suppression with FMF-01. These are deliberately local results: no component behavior is asserted outside its measured range, and the filter trial did not establish overload relief, a new signal or an anomaly.
Horizon 0's calibration-and-evidence software is implemented, but the physical terminated baselines, receiver spur registries and raw Touchstone export campaign remain pending supervised sessions. Until those records are completed and reviewed, the project does not assert absolute field strength, antenna gain or feed-line loss, system noise figure, receiver sensitivity, broadband calibrated power, frequency correction, linearity, image rejection or intermodulation performance. There is still no independent calibrated RF source, and existing marker photographs are not substitutes for reusable complex response traces.
Current operation remains receive-only and supervised. The owner dashboard can launch bounded environmental maps and separate catalog-research cycles, while hardware calibration, review, approval and publication remain distinct human-controlled steps. Raw-IQ model training, demodulation, protocol decoding, automatic emitter identification, automatic anomaly claims and unattended publication remain disabled. A mapped region stays provisional or unresolved until termination, retuning, gain, timing, cross-receiver and source-context controls support a narrower conclusion.