Evaluate the specification that applies to the instrument's operating conditions.

A temperature model records the base accuracy, reference temperature or reference band, valid temperature limits and coefficient form. The author can express supported terms as ppm, a percentage of reading or range, an absolute quantity or an accuracy multiplier, with the applicable function and range identified.

The evaluation retains the observed temperature, its source and time, the coefficient, the resulting contribution and the specification revision. Calibration temperature and autocalibration temperature are separate conditions where the authored specification uses them. The model identifies how each term applies and avoids counting a temperature allowance again when it is already included in the base statement.

A temperature allowance remains an accuracy or uncertainty term unless an authored correction model defines a signed adjustment. It does not silently change the DUT acceptance limit, and the resolver does not extend the coefficient beyond its declared conditions.

Conditional specifications can depend on time since calibration, warm-up duration, autocalibration status, resolution or digits, operating mode, frequency, loading and environmental conditions. A condition may select a different base statement, contribute a numerical modifier or define when a specification is valid.

Standard preserves that distinction. A missing warm-up requirement is not converted into an arbitrary extra uncertainty, and an operating-mode restriction is not treated as a coefficient. The reviewer can inspect the conditions that selected the rule, the modifiers applied and any condition that prevents the specification from being used.

Environmental values can apply to the run or to an individual point. Each observation retains its source, capture time and units, allowing a changing condition to be related to the point measured at that time. Missing, stale or unsuitable values remain identifiable rather than being replaced by one assumed ambient value for the entire job.

An individual instrument's calibration coefficients or certificate-derived corrections retain their own source, validity period, uncertainty, operating conditions and revision. The measurement model states how the correction applies to the result and how its uncertainty contributes. Manufacturer performance limits stay separately identifiable, so an instance correction cannot silently replace a limit or become a second copy of an already included contribution.

Piecewise specifications identify their ranges and the inclusion of each boundary. Selection rules make overlaps, gaps and precedence explicit, and the same instrument, point and conditions resolve to the same rule. A gap or an incompatible condition receives a named outcome instead of a guessed coefficient or hidden extrapolation.

A change to readings, conditions, specification revisions, corrections or calculation settings invalidates the draft results that depend on it and recomputes their analysis. The reviewer can inspect the inputs and rule set used for the new result before release.

Released records preserve their evaluated evidence. A later correction proceeds through controlled amendment with predecessor and successor lineage, keeping the earlier result and the reason for the change available. The laboratory can improve a model without rewriting the history of work already released.