F.14:15 - SoTA-Echoing
F.14 does not import access-control, terminology, credential, or modeling-language taxonomies as FPF ontology. It uses the sources below only where they change the anti-explosion rule.
| Anti-explosion question | Exact source and source-use status | Adoption or rejection in F.14 | Currentness and reopen condition |
|---|---|---|---|
| Why is a system-role-kind label insufficient for authorization? | Rose et al., NIST SP 800-207, Zero Trust Architecture (2020), is a current security-architecture reference that separates a subject’s access to a resource, policy decision, policy administration, and policy enforcement. | Adapt the separation. Keep the kind name, assigned System, request, requested resource and action, policy decision, permission, and Work distinct. Reject authorization, capability, or trust inferred from a system-role-kind label. | Reopen when NIST replaces SP 800-207 or a stronger authorization architecture changes the separation among subject, policy, decision, and enforcement used by this rule. |
| Why should role-like convenience names not replace an explicit policy relation? | Cutler et al., Cedar: A New Language for Expressive, Fast, Safe, and Analyzable Authorization (OOPSLA 2024 extended version), is a current primary policy-language source separating principal, action, resource, context, policy, and authorization decision while supporting role-, attribute-, and relation-based policies. | Adapt only the explicit-policy lesson. Recover the direct policy relation and its participants instead of minting a hybrid system-role kind. Reject importing Cedar entities, schema, or evaluator as FPF ontology. | Reopen if current policy-language practice shows that the explicit boundary between participants and policy no longer prevents the name explosion addressed here. |
| Why must a governed value be recovered before a durable designation or family is minted? | ISO 704:2022 is a current terminology standard connecting objects, concepts, definitions, and designations. | Adopt. Recover the value and use first; then choose no durable name, an existing designation, a local expression, a NameCard, or a public row only at its own trigger. Reject shared spelling as value identity or semantic equivalence. | Reopen when ISO 704 or F.17 and F.18 change the distinction between a value and its designation or the publication threshold used here. |
| Why are credential presentation, status, and relying use different from the governed value? | W3C Verifiable Credentials Data Model v2.0 (2025) is a current W3C Recommendation separating issuer, subject, holder, verifier, credential, presentation, and credential status, and leaving authorization decisions outside the data model. | Adapt. Keep status, evidence, credential, view, verifier action, and relying decision distinct. Reject a suffix, badge, credential view, or dashboard row as a system-role kind, assignment, permission, assurance, or decision. | Reopen when the VC Recommendation or its status family changes the boundaries among presentation, status, and relying use applied in the worked cases. |
SysML is intentionally excluded from the positive SoTA basis and from lineage for this pattern. The official OMG SysML 2.0 specification (September 2025) is recorded only as a rejected-popular comparison: a modeling-language role spelling does not independently establish FPF’s local system-role kind, classification, assignment, capability, permission, Method, or Work. Official status and popularity are not evidence for this anti-explosion question. Reopen that rejection only if demonstrated practice supplies a directly relevant, lower-cost kind-admission and assignment boundary that improves the F.14 cases.