A Property Kind Calculus for Measurement Metrology

2. Requirements🔗

What follows is the fixed set of requirements PropertyKindCalculus is specified to meet. The next section, Why a calculus, not a taxonomy, argues why a calculus is needed where a taxonomy stops, and the chapters after it develop and prove each requirement; this section states them up front, as the axes the design is judged against, with a status table at the end mapping each axis to the declaration that discharges it.

The twenty-eight requirements, R1–R28, fall into nine groups — the section headings of the generated traceability matrix at the end of this section:

  • how kinds are structured, and how quantities characterize objects and named parts of named systems (R1–R3, R19, R22);

  • how operations on kinds are gated (R4–R6);

  • how the kind layer is kept consistent with the coarser dimension and unit layers (R7, R8, R16, R17);

  • how values aggregate over parts (R9);

  • how values are represented — the numeric carrier, vectors under one scalar unit, frames and variance (R10, R11, R20);

  • how classifications are certified, of kinds and of units (R12, R13);

  • how uncertainty is propagated, judged numerically adequate, and covered by intervals (R14, R15, R18);

  • what the discipline costs the author — erasure and rendering (R21, R25); and

  • what evidence a value and a module carry — provenance, audit, the checked measurement model, licensed distribution, and the requirements stated on its quantities (R23–R28).

Below, a group whose axes are best read separately takes one heading per axis. R21–R25 entered the catalogue from the ForPhysLib benchmark's design-neutral MR requirements; the trace between the two numberings runs in both directions in the Requirement validation subsection at the end of this section.

  1. 2.1. Kind structure (R1, R2, R3, R19, R22)
  2. 2.2. Operation gating (R4, R5, R6)
  3. 2.3. Soundness bridges (R7, R8, R16, R17)
  4. 2.4. Aggregation (R9)
  5. 2.5. Representation parametricity (R10)
  6. 2.6. Value representation: vectors and scalar units (R11)
  7. 2.7. Value representation in the structural sense: frames and variance (R20)
  8. 2.8. Verified classification: kind-laws that instantiate at the quantity level (R12)
  9. 2.9. Unit classification (R13)
  10. 2.10. Uncertainty and numerical adequacy (R14, R15, R18)
  11. 2.11. Ergonomics and erasure (R21, R25)
  12. 2.12. Provenance and audit (R23, R24)
  13. 2.13. The measurement model, licensed distribution, and stated requirements (R26–R28)
  14. 2.14. Requirement validation — the ForPhysLib benchmark
  15. 2.15. Out of scope (for now)
  16. 2.16. The requirements at a glance — the traceability matrix
  17. 2.17. What proved is checked to mean