Corrections

A correction should repair the underlying evidence or structured fact, not merely hide a visible symptom in the interface.

What can require correction

Examples include an incorrectly identified model, a source applied to the wrong model scope, a transcription problem, a normalized value that does not match its evidence, or explanatory copy that overstates what is known.

Review before changing a fact

A factual correction should be checked against the relevant source material and provenance before the current value is replaced or adjudicated.

Downstream results must follow

When a corrected fact affects a fit decision, the decision should be recalculated from the corrected source of truth rather than manually patched in the user interface.

Start with the appliance

If you already know the exact model number, return to the ReplaceIQ model analysis tool.

Analyze a model →