Published

Observable Result Fields

vr.wm-xct-025 · wm-xct-025-observable-result-fields

Provide a reusable, format-neutral field group for expressing an observation or measurement result — measurand binding, value, unit and quantity kind, method reference, uncertainty, limits, time anchoring, state, provenance and disclosure class — so that any host record can carry an interpretable, comparable result without forking metrology semantics.

World Models Cross-cutting context XCT.OBS

Bundle → Layer → Finding → Questions Classifiers What it is Why it exists Distinguishing features What robots and AI may and may not do Moral aspects Owners Links to other meta-models Identity and identifiers Direct properties Recognition Capabilities and actions Hazards and failure modes Standards and interfaces Context of use Sources

Bundle → Layer → Finding → Questions Filled

7 bundles · 16 layers · 25 findings · 100 questions

Result Semantics and Host Binding What the reported value is about, how it is shaped, and how the field group attaches to a host record.

Measurand and Observed Property

Identification of the property observed, the entity it belongs to, and the specification detail that makes the value interpretable.

Measurand binding and specification

A value is meaningless without the property it quantifies, the entity that property belongs to, and the conditions under which it was specified; VIM treats a result as quantity values attributed to a measurand plus all other relevant information.

Which observable property, under which specified conditions, does this value quantify? definition
Which feature of interest or sample does the result apply to, and is that the ultimate or a proximate subject? relationship
Is the observed property a quantity, an ordinal property or a nominal property? classification
Below which set of measurand descriptors must the value be rejected as uninterpretable? requirement

Value Representation

The permitted shapes of a result value and how their precision and structure survive storage and exchange.

Value kind and primitive carriage

FHIR carries thirteen value[x] alternatives and SOSA offers both a simple literal result and a structured Result object; a reusable field group must fix which shape a host property uses and how the primitive is carried.

Which value representation is used for this property, and is that choice fixed for the host record type? classification
How is the primitive value carried so that its declared decimal precision survives storage and re-encoding? interoperability
What makes two recorded values of this property the same value for deduplication? identity
Which value representations are prohibited for this property, and on what grounds? constraint

Structured, coded and referenced results

Composite observations (components), bounded results (range, ratio), ordinal and nominal codes, and results too large to embed all need declared structure and ordering.

When one observation yields several correlated components, how are they grouped, coded and ordered? composition
How is a coded or ordinal result bound to its permissible value set and to a defined ordering? classification
How is a result expressed as a range or ratio distinguished from a point value carrying uncertainty? definition
Which reference carries a result held as a large series or waveform outside the record? relationship

Host Attachment

How the field group is mixed into a host record type without collision or ownership ambiguity.

Mixin attachment contract

As a mixin, the group has no standalone record; its attachment point, cardinality, mandatory subset and namespace must be declared per host record type.

Where in the host record is the field group attached, and with what cardinality? composition
Which fields of the group are mandatory for every host, and which are profile-optional? requirement
Which party owns the host record and is therefore accountable for the embedded result fields? ownership
How is a name collision between group fields and pre-existing host fields resolved? exception

Quantity, Unit and Numeric Expression Binding a numeric value to a unit code, a quantity kind and a declared precision, and expressing it canonically without losing the reported form.

Unit and Quantity Kind Binding

Which unit code system is pinned, and how the unit is proved commensurable with the property's quantity kind.

Unit code and code system pinning

FHIR separates a human-readable unit from a system-plus-code pair; UCUM's code space is open, so the code system and its version must be pinned rather than assumed.

Which unit code system is used, and which version of it is pinned to this result? provenance
Is a unit mandatory for this property, and what is recorded when the quantity is dimensionless? constraint
How is a human-readable unit label prevented from being processed as the machine-readable unit code? interoperability
Which authority may approve a unit code that this model will accept? authority

Quantity kind and dimensional consistency

OMS requires the result scale to suit the observed property; QUDT's dimension vectors make that check computable rather than editorial.

Which quantity kind and dimension vector does the observed property require? measurement
What check proves that the recorded unit is commensurable with the declared quantity kind? validation
How are scaled or offset units such as degrees Celsius handled so that arithmetic is not corrupted? constraint
What is recorded when a unit resolves but the property declares no quantity kind? exception

Numeric Expression and Canonical Form

Precision, rounding, and the relationship between the as-reported value and any canonical-unit expression.

Precision, resolution and rounding

FHIR states that decimal precision is significant and must be preserved; resolution of the method is distinct from uncertainty and must not be conflated with it.

How many significant digits are carried, and does trailing-zero precision survive a round trip? quality
What is the declared instrument or method resolution, and how does it differ from the stated uncertainty? measurement
Which rounding rule applies when the value is presented or re-encoded at lower precision? process
Which numeric type limits apply to stored values, including range and decimal versus binary representation? constraint

As-reported value and canonical expression

Comparison across records often needs a canonical unit, but the as-reported value must remain authoritative and the conversion must be attributable.

Is the as-reported value retained unchanged alongside any canonical-unit expression? provenance
Which conversion factor and which registry version were applied to produce the canonical expression? process
Which form is authoritative for comparison and aggregation queries? decision
Under which conditions must a conversion be refused rather than performed? exception

Method, Conditions and Metrological Traceability The procedure that produced the value, the conditions that qualify it, and the chain that makes it traceable and attributable.

Procedure Reference and Conditions

Citation of the method and the subject-specific parameters and influence quantities that qualify the value.

Method and procedure reference

The mixin cites the procedure, algorithm or process chain and its version; authoring and approval of that procedure belong to the method catalogue.

Which procedure, method or algorithm produced this result, and by which identifier is it cited? provenance
Which version or revision of the cited method applies to this specific result? identity
Is the method reference mandatory, and what is recorded when the method is unknown or proprietary? requirement
Which results for the same property may be compared across different methods? interoperability

Conditions and qualifying parameters

SensorThings carries structured parameters on an observation and OMS allows observation parameters; influence quantities that change interpretation must travel with the value.

Which named parameters qualify this result and must accompany it to keep it meaningful? composition
How are influence quantities such as ambient conditions recorded against the value? measurement
Which parameter values are subject-specific overrides of the referenced method's defaults? constraint
How is a required but unrecorded condition flagged on the result? exception

Traceability and Responsible Parties

References that let a consumer judge whether the value is metrologically anchored and who stands behind it.

Traceability and calibration reference

VIM defines traceability as a documented unbroken chain of calibrations; the mixin carries the reference and a validity check, never the calibration process.

To which metrological reference is the value traceable, and through which documented chain? authority
Which calibration record applies, and was it valid at the moment of observation? temporal
What is recorded when no traceability chain exists for the observed property? exception
Which evidence supports the traceability statement without asserting laboratory accreditation? evidence

Observer and accountable organisation

SOSA allows a sensor to be a device, a human or a computational process; a defensible result identifies the observer and the organisation accountable for it.

Which device, sensor or human observer made the observation, and how is it identified? identity
Which organisation is accountable for the reported value? ownership
How is a result distinguished when the observer is an algorithm rather than a physical sensor? classification
What is recorded when the observing device may not be disclosed to a consumer? privacy

Uncertainty, Limits and Qualification How doubt about the value is expressed, how limits and censoring are recorded, and what interpretation context travels with the result.

Uncertainty Statement

Components and their evaluation basis, and the summary statement published with the value.

Uncertainty components and evaluation type

The GUM distinguishes Type A evaluation by statistical analysis of repeated observations from Type B evaluation by other means; each component's distribution and degrees of freedom shape the combined result.

Which uncertainty components contribute to this result, and what does each represent? measurement
Was each component evaluated by statistical analysis of repeated observations or by other means? classification
Which probability distribution and degrees of freedom are assumed for each component? constraint
How is uncertainty judged negligible recorded so that it is not read as unknown? exception

Reported uncertainty and coverage statement

GUM clause 7 requires a reported result to identify whether the stated uncertainty is standard, combined or expanded, and for expanded uncertainty to state the coverage factor and the level of confidence.

Is the reported uncertainty a standard, a combined standard or an expanded uncertainty? definition
Which coverage factor and coverage probability accompany an expanded uncertainty? quality
In which unit is the uncertainty expressed, and is it absolute or relative? constraint
Which uncertainty statement must be present before the result may be published as final? requirement

Limits and Censored Values

Detection and quantitation limits and the recording of values that fall outside a measurable range.

Detection limits and censored values

Regulators require a determined, method-specific detection limit, and FHIR carries a comparator on a Quantity; together they let a below-limit result be recorded without being mistaken for zero.

Which detection or quantitation limit applies to this result, and how was it determined? measurement
How is a value below the limit recorded so that it is not read as zero or as a measured number? constraint
Which comparator operator is carried with a bounded value, and is the bound inclusive? definition
Which handling rule is permitted when censored values are aggregated downstream? process

Interpretation and Reference Context

Flags, quality assessments and reference intervals carried with the value, stopping short of any acceptance decision.

Interpretation flags and reference context

FHIR carries interpretation codes and reference ranges, and SensorThings carries resultQuality; these are context that supports a reader, not a conformity decision, which JCGM 106 treats as a separate discipline.

Which qualitative interpretation flags are attached to the value, and from which code set? classification
Which reference interval or threshold set was applied, and to which population or specification does it belong? relationship
How is an interpretation kept distinct from a formal conformity or acceptance decision? decision
Which quality assessment outcomes are carried with the result and which are only referenced? quality

Temporal Anchoring and Result State When the value applies, when it was produced and ingested, how long it stays valid, and how its record state changes.

Time Anchoring

Separate anchors for the phenomenon, the result and the record, plus validity and aggregation windows.

Phenomenon time, result time and ingestion time

SOSA defines phenomenon time as the time the result applies to the feature of interest and result time as the instant the activity completed; the time the record entered a store is a third, distinct anchor.

When did the observed phenomenon occur, as distinct from when the result was produced? temporal
When was the result record first ingested and made available to consumers? provenance
Which timestamp format, precision and offset are required on each time field? constraint
How are results ordered when phenomenon time and result time disagree? decision

Validity window and aggregation period

SensorThings carries validTime for the period a result is valid; a value may also be a statistic over an interval rather than an instantaneous reading, which changes its meaning entirely.

For which period is the result asserted to remain valid? temporal
Does the value represent an instantaneous reading or a statistic computed over an interval? classification
Which aggregation function and sampling period produced an interval-based value? measurement
What happens to a result once its validity window has lapsed? lifecycle

Result State and Absence

Lifecycle status, supersession of corrected values, and coded absence when no value exists.

Status, amendment and supersession

FHIR defines a status lifecycle from registered and preliminary through final, amended and entered-in-error; a corrected result must link to what it replaces rather than overwrite it.

Which lifecycle status does the result record carry, and which transitions are permitted from it? lifecycle
How is a corrected result linked to the value it replaces without deleting the earlier record? state
Who is authorised to promote a result to final or to amend a finalised value? authority
How long must superseded result versions remain retrievable? retention

Coded absence of a value

FHIR requires a coded dataAbsentReason when no value is present; absence must be distinguishable from zero, from an empty string and from a censored value.

When no value is present, which coded reason explains the absence? definition
How is an absent result distinguished from a value of zero, an empty string and a censored value? constraint
Which absence reasons block downstream use and which are merely informational? validation
Is a record with no value still a valid result record, and under what conditions? requirement

Provenance, Derivation and Evidence Where the value came from, what was applied to it, who asserts it, and what substantiates it.

Derivation and Assertion

Prior values, corrections applied, and the distinction between measuring and asserting a result.

Derivation, correction and assertion

A reported value is often a corrected or derived quantity; PROV-O separates the entity's derivation from the agent to which it is attributed, and VIM notes that a result may be an indication, an uncorrected result or a corrected result.

From which prior values or raw indications was this result derived? provenance
Which corrections were applied, and is the uncorrected indication retained? process
Which agent asserts this result record, as distinct from the agent that performed the measurement? ownership
Which derivation steps must be reproducible for the result to be defensible? evidence

Supporting Evidence

The captured records and budget documents that substantiate a reported value.

Evidence backing the reported value

Raw captures and uncertainty budgets exist as separable objects with their own media, identity and integrity requirements, and are bound to the result by reference and digest.

Which raw or intermediate records substantiate the reported value? evidence
How is an evidence item bound to exactly one result record and verified as unaltered? security
How long must supporting evidence be kept relative to the result itself? retention
Which evidence may be released to a consumer not permitted to see the raw capture? access

Interoperability, Profiling and Disclosure How the field group aligns to external standards, projects into exchange encodings, and carries disclosure constraints.

Alignment and Versioning

Which external standards are mapped, what is claimed, and how code-system change is handled.

Standard alignment and code-system versioning

Alignment to OMS, SOSA/SSN, SensorThings, FHIR, UCUM and QUDT is a mapping rather than a conformance claim; each mapped code system evolves independently and must be version-pinned per record.

Which external standards is this field group mapped to, and is conformance claimed or only mapped? interoperability
How is a code-system version change handled for results already recorded? lifecycle
Where do the aligned standards conflict, and which local rule prevails? constraint
Which profile declares the permitted value kinds, unit systems and uncertainty conventions for a given host? definition

Exchange and Disclosure

Projection into target encodings and the sensitivity constraints that travel with a value.

Projection and round-trip fidelity

The same result must project into OMS, SOSA/SSN, SensorThings and FHIR shapes; each projection loses something (for example SensorThings binds the unit at Datastream level, not per observation) and that loss must be declared rather than discovered.

Which target encodings must the field group project into without semantic loss? interoperability
Which fields are lost or approximated in each projection, and is that loss declared to consumers? quality
How is round-trip equivalence of a projected result verified? validation
Which canonical form is hashed to detect a changed result payload? security

Sensitivity classification and redaction signalling

A value may be more sensitive than its host record; the mixin classifies and signals redaction so that a consumer never mistakes a redacted result for a complete one, while all evaluation stays with the access model.

Which sensitivity classification applies to the value, and does it differ from the host record's classification? privacy
Which parts of the field group may be redacted while leaving the result usable? access
Which legal or contractual basis restricts disclosure of this value? authority
How is a redacted result signalled so that consumers do not treat it as complete? exception

Classifiers Filled

Family: World Models
Category: Cross-cutting context
Entry kind: mixin
Navigation path: NAV.XCT.OBS
Domain: XCT.OBS
Industry: Cross-industry
Tags: observableresultfieldsxct.obs

What it is Filled

This mixin defines the fields that make a reported value interpretable and comparable, and the rules that govern them. It attaches to a host record (registered parent WM-MAT-008 and any other adopting host) and carries references into vocabularies, procedures, devices, samples, access policy and conformity-decision models without reproducing their lifecycles. Storage and interface (JSON, RDF, Markdown, Git, MCP, MongoDB) are projections of these semantics, never their definition.

In scope

Measurand and observed-property binding, including specified conditions and property class (quantity, ordinal, nominal)
Value representation: scalar, coded, ordinal, boolean, range, ratio, component/tuple and reference to an external series
Unit code and code-system binding, quantity kind, dimensional commensurability, canonical expression alongside the as-reported value
Numeric precision, declared resolution and rounding rules
Reference to the procedure or method and the subject-specific parameters and influence quantities that qualify the value
Uncertainty statement: components, evaluation type, distribution, combined standard uncertainty, coverage factor and coverage probability
Detection and quantitation limits, comparator operators and censored values
Phenomenon (event) time, result time and ingestion time, validity window and aggregation period
Result status, supersession and amendment linkage; coded absence of a value
Derivation and assertion lineage, evidence references and digests
Sensitivity classification, redaction signalling and retention class recorded as references

Out of scope

The observation act, sampling activity and their scheduling (owned by an observation/event model)
Sensor, device and platform registration, calibration execution and certificate issuance
Minting, deprecation or maintenance of unit codes, quantity kinds and conversion factor tables
Authoring, validation studies and approval lifecycle of procedures and methods
Feature-of-interest, sample and specimen definition and custody
Numerical propagation of distributions through a measurement model (performed by the producing measurement system)
Conformity assessment, acceptance decisions, guard banding and pass/fail rules
Access evaluation, enforcement, audit-trail storage and physical disposition execution
Time-series, coverage and statistical dataset storage, indexing and compression
Laboratory accreditation, quality-management and competence claims
Alerting, clinical or operational decision support acting on a result

Why it exists Filled

Distinguishing features Filled

A mixin of result fields that make a reported value interpretable: value, unit, quantity kind, uncertainty and completeness.
Differs from the observation act, which records who observed what and when.
Differs from sensor and calibration records, which it references.
Records absent, censored and below-limit values explicitly instead of as blanks or zeros.

What robots and AI may and may not do Filled

Must not

Drop the unit or uncertainty when projecting a result.
Replace a censored or missing value with a number.
Compare results measured by incompatible methods as if equivalent.
Change a reported value instead of superseding it with a correction.
Invent precision by adding digits after conversion.

Only with a human decision

Accepting a result for a safety or clinical decision when its comparability is doubtful.
Approving a correction to a reported value.

May

Validate a unit against the quantity kind.
Convert to a canonical unit expression while keeping the reported one.
Assemble an uncertainty statement and check it is complete.
Assess comparability of two results before comparing them.

Moral aspects Filled

Results drive clinical, safety and environmental decisions; misread units can harm people.
Missing uncertainty can make weak results look certain.
Results about people, such as health measurements, are sensitive data.

Who is affected

People whose samples or bodies were measured
Decision makers who rely on results
Laboratories and observers

Owners Filled

Steward

The adopting Dimension names one accountable steward for the field group and one owner for each host binding profile, recorded with contact and escalation paths.

Roles

Field-group steward: Maintains the field definitions, mandatory sets and canonicalisation rules; Owns the alignment and conflict register and decides breaking-change classification; Approves new binding profiles and their versions
Host record owner: Accountable for correct population of the group within their record type; Selects and keeps current the binding profile for that record type; Resolves collisions between group fields and host-native fields
Measurement authority: Approves uncertainty conventions, coverage factor defaults and negligibility thresholds; Approves the pinned unit code system, canonical units and traceability claim rules; Reviews detection and quantitation limit sources for each property
Result asserter: Submits and finalises result records and is accountable for the reported value; States amendment reasons and supersession links for corrections; Declares absence, censoring and negligible-uncertainty cases explicitly
Interoperability reviewer: Maintains projection mappings and the declared loss report per target; Reviews code-system version changes and their effect on existing records; Runs and records round-trip equivalence tests

Links to other meta-models Filled

child

WM-MAT-008 (registered parent model) - This field group is registered as a child of WM-MAT-008 and supplies the result payload fields its property records carry; the parent owns the record type, its subject and its lifecycle.

composes

Host record models adopting the field group (any model carrying observed or measured properties) - Hosts embed the group under a versioned binding profile; the host owns the record, its identity and its access class, while this model owns only the field semantics and their validation rules.

references

Unit and quantity-kind vocabulary registry (UCUM, QUDT, UN/CEFACT Rec 20) - Supplies unit codes, quantity kinds, dimension vectors and conversion factors. This model pins a code-system version and cites codes; it never mints, deprecates or maintains them and holds no local copy of the conversion tables.
Observed-property and observable vocabulary registry - Supplies the observed-property codes and their expected quantity kinds and permissible value sets; this model carries the code, the code system and the version pin only.
Procedure and method catalogue model - Supplies method identity, versions, defaults and validation studies. This model carries the reference, the version and subject-specific parameter overrides, and never reproduces method authoring or approval lifecycle.
Sensor, device and calibration record model - Supplies observer identity and calibration records. This model carries the reference and a validity-at-observation check; device lifecycle, calibration execution and certificate issuance remain with the target.
Feature-of-interest, sample and specimen model - Anchors what the value is about. This model carries the reference and a proximate-versus-ultimate subject flag, not sample identity, hierarchy or custody.
Agent, party and organisation identity model - Supplies observer, asserter and accountable-organisation identities; this model carries references and role codes only.
Access policy, audit and retention model of the adopting Dimension - Receives the sensitivity classification, retention class and policy reference recorded here and owns policy evaluation, enforcement, audit-trail storage and disposition execution. This model supplies audit event content but never evaluates, enforces or stores the trail.
Time-series and statistical dataset model - Holds series and waveform payloads that a result may point to; this model carries the series reference plus the sampling and aggregation descriptors needed to interpret it.
Conformity assessment and decision-rule model - Consumes values, uncertainties and limits to make acceptance decisions under a decision rule; this model supplies inputs and interpretation context and never records the conformity decision or its guard bands.

aligned

JCGM measurement uncertainty framework (GUM JCGM 100:2008 and VIM JCGM 200:2012) - Field names and semantics for measurand, measured quantity value, standard, combined and expanded uncertainty, evaluation type, coverage factor and coverage probability follow GUM and VIM terminology; this is a documented mapping, not a conformance claim.
OGC/ISO Observations, Measurements and Samples and W3C SOSA/SSN - Observed property, procedure, feature of interest, phenomenon time and result time map to OMS and SOSA constructs so that a hosted result can be exposed as an observation result without re-modelling the observation act.
HL7 FHIR Observation and Quantity datatype - Value polymorphism, comparator, absence coding, interpretation, component structure and status lifecycle map to FHIR equivalents for cross-domain exchange; FHIR resource-level security and audit semantics are explicitly not adopted.
W3C PROV-O provenance vocabulary - Derivation, attribution and generation-time fields map to wasDerivedFrom, wasAttributedTo, wasAssociatedWith and generatedAtTime; the provenance graph store and its query semantics remain external.

neighbor

Observation act / event model (sosa:Observation, OMS Observation) - The act carries exactly one result; this mixin defines the result payload's fields and their interpretation rules. It does not own the act, its sampling features, its stimulus or its scheduling.
Sensor, device and calibration record model - This mixin carries an observer/device reference and a calibration record reference plus a validity-at-observation check. Device lifecycle, calibration execution and certificate issuance stay with that model.
Unit and quantity-kind vocabulary registry (UCUM, QUDT, UN/CEFACT Rec 20) - This mixin pins a code system version and cites codes, quantity kinds and conversion factors by reference. Code minting, deprecation and factor maintenance are the registry's.
Procedure and method catalogue - This mixin carries a method identifier, version and subject-specific parameter overrides. Method authoring, validation studies and approval remain in the catalogue.
Conformity assessment / decision-rule model - Interpretation flags and a reference to an applied reference interval are carried here as context; acceptance, pass/fail and guard-band decisions and their records belong to the decision model.
Feature-of-interest, sample and specimen model - Referenced only, to anchor what the value is about; sample identity, hierarchy and custody are not modelled here.
Access policy, audit and retention model of the adopting Dimension - As FHIR separates Observation from its security and audit resources, this mixin records a sensitivity class, a policy reference and a retention class only; evaluation, enforcement, audit-trail storage and disposition execution are external.
Time-series / statistical dataset model (SensorThings Datastream, FHIR SampledData) - When the result is a large series the mixin holds a reference plus the fields needed to interpret it; series storage, windowing and aggregation infrastructure are external.

parent

WM-MAT-008

What else AI and robots need to interact with it Filled

Sources Filled

OGC Abstract Specification Topic 20: Observations, Measurements and Samples (OGC 20-082r4) / ISO 19156:2023 - Open Geospatial Consortium and ISO/TC 211
Semantic Sensor Network Ontology (SSN/SOSA) - World Wide Web Consortium (W3C)
JCGM 100:2008 Evaluation of measurement data — Guide to the expression of uncertainty in measurement (GUM 1995 with minor corrections) - Joint Committee for Guides in Metrology / BIPM
JCGM 200:2012 International Vocabulary of Metrology — Basic and general concepts and associated terms (VIM), clause 2.9 measurement result - Joint Committee for Guides in Metrology / BIPM
The Unified Code for Units of Measure (UCUM) - Regenstrief Institute, Inc. and the UCUM Organization
HL7 FHIR Release 5 — Observation resource - Health Level Seven International
HL7 FHIR Release 5 — Data Types (Quantity, Range, Ratio, SampledData) - Health Level Seven International
QUDT — Quantities, Units, Dimensions and Data Types Ontologies - QUDT.org
OGC SensorThings API Part 1: Sensing Version 1.1 (OGC 18-088) - Open Geospatial Consortium
PROV-O: The PROV Ontology - World Wide Web Consortium (W3C)
The International System of Units (SI Brochure) - Bureau International des Poids et Mesures (BIPM)
JCGM Publications: Guides in Metrology (JCGM 100, 101, 102, GUM-1:2023, GUM-6, JCGM 106:2012, JCGM 200) - Joint Committee for Guides in Metrology / BIPM
Procedures for Detection and Quantitation Documents (Definition and Procedure for the Determination of the Method Detection Limit, Revision 2, EPA 821-R-16-006) - United States Environmental Protection Agency

Open questions

Close the declared limits-and-censoring gap: verify cross-domain detection and quantitation limit determination against primary text from IUPAC, ISO 11843 and clinical limit-of-blank / detection / quantitation conventions, so the semantics do not rest solely on a United States Clean Water Act procedure.
Close the declared measurement-scale-type gap: find a verifiable cross-domain primary source for scale type (nominal, ordinal, interval, ratio) and for permitted arithmetic on ordinal codes; the nominal-property counterexample in the pack's own adversarial check depends on this provisional field.
Obtain licensed access to ISO/IEC 17025 clause 7.8 to confirm that the deliberate non-claim about test-report conformance remains correct and that no mandatory reporting element is silently missing from the mandatory descriptor set.
Decide whether correlation and covariance between uncertainty components or output quantities belong in this field group at all, or wholly in the producing measurement system, and record the decision rather than leaving multivariate results as components with no covariance structure.
Re-verify whether SSN/SOSA has moved beyond the 2017 Recommendation in a way that changes resultTime typing, since the model's divergence on offset-free timestamps is stated as a conflict against that specific version.
Resolve the two missing functions in a later revision: an explicit code-system re-pin migration function required by crud.update and compatibility_rules, and a redaction-emission function required by crud.read, neither of which could be added under single-provider rules.
Determine whether a normative rounding algorithm can be sourced at all, or whether the current 'named but not mandated' position is permanent; the pack leaves rounding unresolved for lack of a verified cross-domain source while still requiring trailing-zero precision to survive round trips.
Numerical propagation of distributions through a measurement model (GUM Supplements 1 and 2) is deliberately excluded; only the resulting statement is carried.
Correlation and covariance between components or between multiple output quantities are not modelled, so multivariate results are carried as components without a covariance structure.
Spatial reference semantics for geometry-valued results (CRS, axis order, positional uncertainty) are not modelled and would need a geospatial sibling model.
Reference material and certified value semantics are referenced through traceability but not modelled as fields.
Rounding and significant-figure rules are named but no normative rounding algorithm is mandated, because no single cross-domain primary source was verified.
Currency, monetary and legal-metrology-specific value semantics are not addressed.
Localisation of decimal separators, digit grouping and unit rendering for human display is left to presentation layers.

Machine files

Card as JSON Specification AGENTS.md Source Card backlog vercy.card/1

Provenance

world-models research · reviewable-draft

Built from: models/wm-xct-025-observable-result-fields/spec.yaml, ver-cy/world-models/card-supplements/wm-xct-025-observable-result-fields.json