Where the idea comes from
Source-backed. The mechanism is described in the cited primary source. Its performance in another environment requires a separate test.
Skufosaur method. A proposed way to investigate a problem or compare options. Results are reported after the check runs.
When I call a result verified
Record the original task, environment versions, constraints, baseline, candidate and acceptance criteria. Include repeat results and artefacts that let someone else run the check. An article edit date records the edit.
What I measure separately
- Result quality and compliance with the task constraints.
- Time to a usable result and work across all branches.
- Reported usage, estimated cost and actual charges.
- Completeness, adapter errors and unknown values.
How I compare alternatives
Change one approach at a time and hold conditions and acceptance criteria constant. Repeated runs show variation. Count discarded attempts in the cost.
Sources and corrections
Primary sources show what their authors claimed. Analysis explains where a claim applies and what still needs checking. A correction states what changed and which conclusions it affects.
Basis for the method
The internal Master Codex v5.3 distinguishes observation, interpretation, experiment and demonstrated effect. A partial or blocked check remains partial or blocked.