How a lesson earns publication

Every lesson in this library is drafted by an automated pipeline and then reviewed against its sources by a named human editor before it is published. The two steps are deliberately separate: a draft is a candidate, not a publication. Nothing reaches the site on the strength of having been generated.

Who is accountable

Bob Smith, Editor, owns these standards, reviews lessons against their sources, and makes the final decision to publish. The method below is the standard he holds each lesson to. Accountability for what appears here is human, not automated.

The two gates

A candidate lesson passes two independent checks before it can be published. The first is structural: a validator enforces the rules every lesson must satisfy — length, a working vocabulary glossary, five questions, and a specific failure-mode rationale for every wrong answer. The second is substantive: each candidate is scored across five dimensions —

• Clarity — is the explanation legible to its intended reader without dumbing the mechanism down?
• Accuracy — is it correct, and does it say what it cannot establish rather than overclaiming?
• Pedagogical depth — does it teach the real mechanism, not a memorable but misleading cartoon of it?
• Distractor quality — does each wrong answer name a specific, plausible way of being wrong, so the quiz teaches?
• Guardrail adherence — does it respect the editorial rules specific to its subject?

A lesson that scores well enters a reserve; only lessons from the reserve are published. Borderline lessons are routed for human review rather than appearing automatically. A rejection is treated as a successful outcome — a mediocre lesson published is not.

Corrections

The library is an early and growing version, and errors are possible. Corrections are taken seriously and reviewed by the editor; a dedicated contact channel for reporting them is being established and will be listed here.