Status & coverage
A live account of what is built, what is verified, and whether the full canon is worth committing to. Every number below traces to a gate that either passes or blocks the build.
Quality gates
Gates run in order. A downstream gate is meaningless until every gate above it is green, so the ledger blocks rather than warns.
| G0 |
Schema freeze
derivation record fields fixed: id · statement · assumptions · steps · result · used_by
|
pass |
| G1 |
Units — dimensional check parses
every result equation reduces to consistent SI dimensions
|
pass |
| G2 |
Numeric — verified to 4 s.f.
worked value reproduced from first principles, agrees to four significant figures
|
pass |
| G3 |
Provenance — authored independently
each derivation written from the physics, not transcribed; sources cited for cross-check only
|
pass |
| G4 |
Links — every used_by resolves, no orphans
each lecture cite points to a live derivation; each derivation is reachable from at least one unit
|
pass |
Cost probe
Measuring the real cost of one page
These 20 derivations are not a sample of the canon — they are a stopwatch. Each is built end-to-end through all four gates so we record actual hours-per-page before committing to the full ~410-page canon.
The pilot spans three units and eight threads so the average reflects hard pages, not just the easy ones.
What the mean buys us
under 3h mean two-year project viable The full canon is affordable at a sustainable pace; proceed to seed the remaining units.
over 8h mean cut to a 150-derivation core The canon as scoped is too expensive; drop to a tighter core that still spans every thread.
Between 3h and 8h the call is a judgement, not a rule: re-scope per-thread, or trade rigour tier depth against coverage. This band is where the probe earns its keep.
Permalink · see the method, the derivation vault, or a worked example at D-001.