Keep separate runs separate. Promote only the result you intend to use.
Each set keeps its own readings, unit and computed snapshot. Changing raw readings invalidates only that set's prior summary.
Start with measurement context, then work through repeated readings, averages, conversions, error, uncertainty and decision math in one connected reference session. A working-value register carries results between core tools while session receipts keep the arithmetic handoff inspectable. Nothing here becomes controlled calibration evidence until it enters the separately governed record workflow.
The Atlas configures the quantity tool and asks for the conditions that make the measurand sufficiently described. Completeness is descriptive, never a validation claim. Selecting another discipline or quantity starts a fresh public context draft.
Pressure transmitters, gauges, calibrators, vacuum instruments and reference systems.
This public Workbench describes context only. It does not authenticate an instrument, bind a controlled procedure, create evidence, or authorize release.
Linear unit conversion is available. Mode, medium, reference basis and correction terms remain separate measurement context.
Build multiple reading sets, promote the statistic you actually need, chain it through bounded computation bays, and preserve a machine-readable arithmetic trail without turning exploratory work into controlled calibration evidence.
Each set keeps its own readings, unit and computed snapshot. Changing raw readings invalidates only that set's prior summary.
Add or subtract in the carried unit. Multiply or divide only by a dimensionless scalar. The workstation preserves the unit instead of inventing dimensional semantics.
Comparison is same-unit only. If two runs carry different units, convert explicitly first; the machine will not silently normalize them.
Declared resolution drives half-even reporting. Quantization uncertainty is calculated only under the explicit uniform ±resolution/2 assumption.
Promote the active set to Type A only when the repeated observations actually support that interpretation. Effective degrees of freedom remains a separate uncertainty-model calculation.
Reading set 1 · kPa
Full official decimal SI prefix range. Symbols are case-sensitive: m = milli; M = mega.
Quantity family locked to the Atlas context. Change the configured quantity above instead of silently crossing physical quantities inside the computation session.
Enter standard uncertainties. Correlated contributors require a covariance model and are intentionally blocked from this simplified reference model.
Each supported operation records inputs, deterministic model/version, context snapshot and outputs. Export includes the reading sets and current register, but it still carries reference-only authority.
Computing fingerprint…Keep declared resolution, count terms and count-derived arithmetic estimates as separate concepts with explicit provenance.
Open range command surface →Vendor-neutral instrument discovery, identity and reading architecture across approved protocol and device profiles.
Inspect connectivity architecture →Intervals, drift and stability, due/recall, procedures, certificates and integration governance.
Inspect operations architecture →