3 min read

Reference — read, interact, and evaluate a published standard

WHO THIS IS FOR

The Reference pillar is for Readers — anyone who needs to understand, test, or evaluate against a published standard: engineers checking product fit, test labs running standardised procedures, conformity assessors verifying requirements, students learning the domain, compliance researchers comparing standards. See Readers.

Reference is the consumption side of Primmel. Once a publisher has released a reference .prl, anyone can load it into a tool and do three kinds of work:

  • Read — navigate the standard’s structure: provisions, processes, data classes, references. Find what a clause requires, where it lives, what it traces back to.
  • Interact — exercise the standard dynamically: run the test processes it declares, follow its process flows, evaluate its measurement expressions against sample inputs.
  • Evaluate — check whether a product, process, or dataset meets the standard: dimensional checks, formula evaluation, applicable-requirement checks.

This is fundamentally different from the Application side. Readers don’t maintain an implementation. They don’t produce long-lived evidence. They consume the published standard to answer a question: does this thing meet the requirement?

Output artifacts

Reference use produces transient verdicts, not standing artifacts:

Activity Typical output
Read A navigated view of the standard (in a browser/editor)
Interact A test run — executed process flows, evaluated formulas
Evaluate A pass/fail verdict for a specific product, process, or dataset

These outputs may be saved as test reports or evaluation records, but they are not part of the Primmel artifact set the way .prl, .prm, and .pws are. They are uses of the published reference.

What Readers consume

Artifact Purpose
Published reference .prl The standard itself — provisions, processes, data classes, references
Optional .prd Source clause text, for cross-checking provisions against the underlying standard document

Tools Readers use

  • Spec browsers — navigate the model: tree view, full-text search, cross-reference jumps, dependency graphs.
  • Test harnesses — execute the standard’s test processes against sample or real inputs; produce pass/fail verdicts per step.
  • Formula evaluators — run the standard’s validate_measurement expressions and conditions against provided values.
  • Dimensional checkers — verify whether a product’s dimensions, weights, or other measurable properties fall within the ranges the standard declares.
  • Applicability checkers — given a context (product type, use case, jurisdiction), determine which provisions apply and which do not.

What Readers do not do

  • They do not maintain an implementation .prl of their own operations. That’s the Implement pillar.
  • They do not produce .pws/ evidence of running operations. That’s the Operate pillar.
  • They do not audit implementations against the standard. That’s the Audit pillar — which uses reader-style navigation as one of its capabilities.

Where to see it in action

  • The published reference .prl in the implementation package (ocs-standard.prd alongside the OCS#-aliased declarations in acme-coffee-programme.prl) is exactly what a Reader would load into a tool to navigate, test, or evaluate against.
  • Compliance and measurement example — the gateway edges with condition expressions and the validate_measurement blocks are precisely what a Reader’s formula evaluator would run.

After reading

A Reader’s verdict may feed into other pillars: a conformity assessor’s evaluation can become the basis for a certification (which a Publisher then records); a test lab’s run can become evidence a Regulator reviews. But the Reader activity itself stops at the verdict — the verdict’s downstream use is someone else’s pillar.