Every correction made to a published excavation is logged and stays visible. The logs are append-only: an entry is never deleted or rewritten, only added to.
2026-10-02 · Gulf of Tonkin Incident: The August 4, 1964 Attack Never Happened, NSA's Own Study Found CORRECTION. In this dig's timeline, the 24 Jun 1970 Senate repeal vote (event g9) listed as a source the Voteview page for the 7 Aug 1964 Senate vote (roll call 481), a different vote. Merging that event with the Johnson-era votes dig's copy exposed it.
2026-10-02 · Gulf of Tonkin Incident: The August 4, 1964 Attack Never Happened, NSA's Own Study Found CORRECTION to the "90 percent" figure. Earlier text said "almost 90 percent of the intercepts that would have contradicted the attack story". The study says the withheld information "constituted nearly 90 percent of all available SIGINT" on 4 August, that 122 relevant products were reviewed, that NSA's official version used 15 published reports, and that the six products used in the chronology are "a bit more than 10 percent of the total available".
2026-10-02 · Flydubai FZ1073: Saudi ministry says co-pilot assaulted captain; no evidence of staging found CORRECTION TO L-13: it said Flightradar24's '14,125 ft in 29 seconds' fits the peak rate. It does not fit: 14,125 ft in 29 s is 29,200 ft/min on average, nearly the whole peak rate with almost no time to build it up, which would need an implausibly abrupt change; CNN's 17,400 ft in just under two minutes averages 8,700 ft/min. The 29 s figure is probably a sampling artefact or measures a different slice; treat as UNRESOLVED until the raw ADS-B CSV is read.
2026-10-02 · Eikon Basilike (1649): Milton Charged Charles I's Prayer Was Copied Word for Word From Sidney CORRECTION TO THE BOUNTY PAGE: the text read is the 1650 'second time and much enlarg'd' edition, not the 1649 first edition the claim refers to; the claim is kept established (the charge is documented) with the edition recorded and a next step to read the 1649 edition. (2) CORRECTION TO THE BOUNTY PAGE: TCP A38258, the text the test uses, is catalogued as 'First edition, second issue' (Madan: 'probably about 4 February'), whereas the page calls it 'the first issue'. Claim eikon-text-tested-is-second-issue records this; it does not change the test, since the text is the one tested.
2026-10-02 · Casket Letters (Mary Queen of Scots, 1567): No Originals Survive, Only Copies and Translations CORRECTION to L-01: L-01 says the drafted headline has status 'review'. That was wrong. Headline rule 2 allows only a claim that is established or refuted at HIGH confidence; casket-originals-lost is established but capped at confidence 'provisional' (an absence anchor), so the headline cannot meet the rule.
2026-10-02 · Casket Letters (Mary Queen of Scots, 1567): No Originals Survive, Only Copies and Translations CORRECTION of my own earlier note: the log entry for the nodes pilot called the Westminster conference dates (7-14 Dec 1568 in the overview and 9-14 Dec in the detail panel) an unreconciled inconsistency. The node texts show they describe different parts of one process: the sessions began on 7 and 8 Dec, Morton exhibited his declaration on 9 Dec when the collation began, and the minute is of 14 Dec. The collation node is now linked as part_of the conference node (established), not as a possible same-event.
2026-10-01 · Gulf of Tonkin Incident: The August 4, 1964 Attack Never Happened, NSA's Own Study Found CORRECTION to the first draft: it said the NSA study left "error or design" open on intent. The study as reported by the press found deliberate skewing by mid-level NSA officers to cover earlier errors, and found that top NSA and defense officials and President Johnson neither knew nor condoned it. The single "deliberate fabrication" claim was therefore split into tonkin-midlevel-deliberate-skew (proposed) and tonkin-senior-knowledge (contested).