A report can show a measured value equal to the printed limit while marking the test as failed. This may be a legitimate consequence of rounding, or it may reveal inconsistent comparison logic. Define how raw values, derived values, limits, and display formatting interact before the ambiguity reaches production review.
NIST’s guidance on expressing measured quantities discusses the meaning conveyed by numerical representation. Displayed digits communicate information, but adding digits does not improve the underlying measurement. Keep presentation separate from the approved acceptance method.
Document the comparison rule
State which numerical value is compared with which limit, whether boundaries are inclusive, and where any permitted rounding occurs. Do not let a spreadsheet’s default formatting decide this rule. If the acceptance specification requires a particular rounding procedure, implement that procedure explicitly and verify it.
Preserve sufficient numerical information for the calculation and later investigation. Avoid rounding at several intermediate stages merely to make temporary values look tidy. Repeated rounding can change a derived result even when the final report uses a reasonable number of digits.
Exercise the boundary deliberately
| Test case | Behavior to verify |
|---|---|
| Value comfortably inside the limit | Expected pass and clear display |
| Value exactly at the comparison boundary | Documented inclusive or exclusive behavior |
| Values immediately on either side | Consistent comparison despite similar rounded displays |
| Converted or calculated quantity | Correct units and approved rounding stage |
Use controlled software test inputs for these cases. They check calculation and reporting behavior; they do not replace validation of the physical measurement or justify narrowing a product tolerance.
Make the report explainable
- Display the unit beside the measurement and limits.
- Use enough precision to make near-boundary results interpretable.
- Retain the underlying comparison value in the detailed record.
- Identify the rule or recipe version that produced the decision.
- Keep uncertainty or guardband policy distinct from display rounding.
When a report appears contradictory, inspect the stored values and comparison implementation before overriding the result. An operator should not have to guess whether a visually equal value ought to pass. Include the same boundary cases in exported spreadsheets and database reports, because a correct station decision can become misleading when a downstream report recalculates status from formatted text instead of the approved numerical record.
Resolve the issue by correcting the rule implementation or clarifying the report, as appropriate. Retain evidence for affected historical records. The measurement-unit review addresses another common source of disagreement between stored values and displayed results.
Move from reading to a build-ready project
Use the fabrication and assembly checklists to prepare drawings, production files and test requirements.
Open the DFM checklist