Chronology
The chronology is built from the documents as they land — every dated event the analysis extracted, merged across documents and put in order. Each row shows how many documents it came from.
It is also what the case summary, research checklist and theory stages read, so a wrong row is not cosmetic.
Correcting a row
Correct this event replaces the extracted wording with yours.
A row you touch stops being the engine’s. Correcting, striking or freezing a row takes it off the engine for good — no later run rewrites it.
The fields
Date — as the record puts it. Early 2020 is a real answer.
Sorts under — optional. Where a fuzzy date belongs in the order. Leave it empty and the row sorts last.
What happened — one sentence.
Who was involved — names, separated by commas.
Adding an event by hand
Add event is for what no document records — the call, the meeting with no minutes. It is yours from the start; no run will ever touch it. Rows added this way are marked Added by hand.
Striking and restoring
Strike from the chronology removes a row from the timeline without deleting it. Struck rows are hidden; Show struck rows brings them back, and Put back restores one.
A row the documents no longer support is marked No longer in the documents rather than vanishing.
Freezing a date
Freeze this date pins a row so runs will not rewrite it — useful where the extractor keeps re-deriving a date you have already settled. Hand back to the engine unfreezes it, and a later run may then rewrite it.
Chronology versus facts
A chronology event happened on a date. A fact is a proposition with no date, handed to every stage of every run. Use the one that fits what you are recording.