Starlight Ranking Compilation and Revision Discipline
Everyone can be assessed with the same measure; the frontier, Top lists, and awards derive from the same person ledger.
Explains how the ranking admits, orders, and revises entries; earlier ranking rules remain available in the version history.
The current Starlight ranking uses people with migrated outcome ledgers as its public point-score scope, replaying the same outcome ledger, formula, and global budget. What is published is a reviewable research point: each outcome retains M, a, sources, gaps, and sharing relationships. Two displayed decimals support replay, not evidence precise to the hundredth or automatic certification.
Unmigrated people retain identity and contribution introductions without legacy scores, ranks, or tiers. Important new evidence, an identity or attribution error, an outcome-boundary change, or a method upgrade may trigger a coordinated revision of the affected scope; older algorithms and differences enter the version history rather than a parallel current ranking. Fame, Top labels, and awards cannot determine M, a, or who is eligible to be assessed.