Before Bitcoin
Calibration
An earlier software project supplied a hypothesis about specifications and executable gates. It is context, not evidence for this case.
05 / Origin
An artifact history, not an effort study.
These non-equivalent commit and artifact spans describe what appeared in the repository and when. They do not measure labor, productivity, agent capability, or causal effects.
20 Apr → 4 May 2026
Formalization
A 30,549-line assertion intermediate language was recorded at the end of a fourteen-calendar-day span. No running node existed at that endpoint.
3 → 23 May 2026
Process
A twenty-calendar-day span produced coordination and provenance infrastructure before the later product artifact.
23 May → 8 Jun 2026
Product track
Java product-track commits appear across a sixteen-calendar-day span. The timestamps describe repository history, not effort or discovery cost.
5 Jun 2026 · 12:38 → 15:56
Zig interval
Commit timestamps delimit 3 h 17 min 57 s from an empty scaffold to a 50k shakedown. The task is not equivalent to the earlier spans and the interval is not an effort log.
17 Jun 2026
Frozen snapshot
Twelve ports had corpus and strict 5k proofs; nine had canonical clean long-run lanes. No port had an empty-state-to-tip proof.
August 2026
Diagnostic supplement
A separate deposit preserves bounded pure-Mojo evidence through 140,234. It is post-snapshot packaging, noncanonical, and not an equivalent end-to-end replication.
Historical archive
The early story survives as provenance—not authority.
The ten pre-release essays preserve what the project believed while it was still inside the tunnel. They are useful context and explicitly superseded where the public evidence changed.
Read the field notes →