FPF C.2.1, A.22, C.30.AD, and C.30.ASV multi-view architecture discipline | FPF separates actual subject relations, selected A.22 structure, ArchitectureRelation, bounded claim content, Description episteme, viewpoint, structural view, representation, publication, correspondence, grounding, and source return. | Ask whether the actual relation obtains or the content is modal, then choose the next architecture move before opening heavier description and view records. When architecture-description or traceability use is current, recover correspondence, source pins, description-reliance relations, and source-return conditions. | A tool, notation, model-use structure, view, description, file, list, or publication creates none of the subject-side relation, structure, or truth facts by form. |
| SEI views-and-beyond lineage and multi-view practice | Keep module, component-and-connector, runtime interaction, allocation, and placement as separate view pressures. | Do not reduce architecture to module structure or interface relation; use C.30.ASV to test structural-view claims. | View taxonomies are lineage and comparison support, not a second FPF ontology. |
| Sapunov, Theory of Code Space, arXiv:2603.00601v4 | The benchmark studies architectural belief maps during budgeted codebase exploration under partial observability; reported benefits of active exploration and retained maps vary by model. | Use a generated map as a bounded source for inspecting architecture. Recover the actual relation meanings, supporting material and any unknown region that matters to the intended use. | The benchmark does not establish the adequacy of an architecture or authorize a code change. |
| Holon-architecture law-like constraint set from the architecture source | Adopt Conway and mirroring as transformer-transformed correspondence pressure through C.32.CONWAY; use other law-like architecture lines only as recognition pressure for selected structures and architecture characteristics. | For Conway or mirroring, recover transformer holon, transformed holon, changing relation, selected structures, affected characteristics, candidate gain, and candidate loss. For other law-like pressure, identify the selected structure and characteristic, then use the pattern that defines or tests the architecture, relation, measurement, selected-set, or decision claim. | A law-like slogan does not by itself establish architecture adequacy, a decision, evidence sufficiency, assurance, or gate passage. |
| GonzoML neural-network architecture corpus as source example for general architecture-operation language | Adopt practitioner architecture-operation language as general architecture material: structural substitution, relation retargeting, dataflow change, path-selection and gating, memory and cache placement, block and layer substitutions, MoE expert-selection, pruning, distillation, NAS, ablation, and compute, memory, and latency tradeoffs. | Keep source labels as source labels through C.30.STRAT; after recovery, use the language for architecture-description and architecture-view recognition, transformation-flow-structure source relation, module-and-interface repair, scale characterization, candidate move guidance, and decision-context fields. | Neural-network labels, benchmarks, ablations, pruning masks, search outputs, or distillation success do not become FPF ontology, architecture decision, evidence sufficiency, gate passage, assurance, or architecture adequacy by themselves. |
| Platform-engineering, MOSA, and open-systems practice | Adapt open-interface, platform extension-rule, substitution-policy, and conformance-expectation pressure as local architecture boundary discipline. | For an open-interface or platform claim, name the local structure, interface, variation point, substitution policy, pattern for any conformance-evidence claim, migration boundary, update channel, and hardening boundary that change action. | Platform design depends on project, organization, time, and place; there is no universal platform maturity scale, and an open label is not proof. |
| ADR and architecture-knowledge-management practice | Adopt decision-memory pressure only as a project-side decision concern handled outside C.30. | Treat ADR-like material as a publication or decision-description source relation until an architecture decision claim is being made. | ADR is not the project decision itself and not a source of release authorization. |