Research record
Research record
Exact target
ID, version, domain, assumptions, quantifiers, conclusion.
Attempts
Keep original outputs in separate files. For each argument: prerequisites, actual calculation, remaining obligation.
Evidence
Exact proof, source locator, counterexample, or actual tool output. Distinguish proposed checks from completed checks.
Admission
Which statement is supported, by what evidence, for which version? Which dependencies remain assumptions?
Revision
What changed? Which conclusions survive, require rechecking, or lose support?
Next experiment
One action, its predicted outcome, and what a contrary result would teach us.
