Repository navigation
intro
This chapter expresses Prism's conceptual model in Object-Process Methodology (OPM) per the normative source:
ISO/PAS 19450:2015 — Automation systems and integration — Object-Process Methodology, First edition, 2015-12-15 (ICS 25.040.40, 35.240.50). Advanced to ISO 19450:2024 (Second edition, 2024-04). Published by the International Organization for Standardization. The 2015 edition (Public Available Specification, ISO/PAS) is the source the wiki's pinned OPL EBNF transcription at
tools/iso-19450-opl.ebnftargets; the 2024 edition advances the same content to a full ISO standard with the same Annex A formal-syntax structure.
Supplementary sources corroborating this chapter's OPM transcription:
- Dori, D. (2002). "From Object-Process Diagrams to a Natural Object-Process Language." In: A.B. Pidduck et al. (Eds.): CAiSE 2002, LNCS 2348, pp. 221–238. Springer-Verlag. Establishes the canonical OPL production rules in their original R-numbered form (R1–R20+); predates the ISO standard.
- Dori, D. (2016). Model-Based Systems Engineering with OPM and SysML. Springer New York. ISBN 978-1-4939-3294-8. Chapter 11 ("Object-Process Language: The Text") aligns with Annex A of the standard.
- ISO 19450:2015 free preview (Foreword, Introduction, Clauses 1–3) — confirms term vocabulary and the bimodal-equivalence principle V7 enforces. Available at iso.org's preview viewer.
Per the standard's Foreword, compliance may involve the use of a patent concerning OPM as a modelling system (Clauses 6–14). The patent holder, Prof. Dov Dori (Technion — Israel Institute of Technology), has assured ISO of willingness to license under reasonable and non-discriminatory terms.
OPM specifies a bimodal representation: each conceptual statement appears as both an Object-Process Diagram (OPD) — the visual rendering — and an Object-Process Language (OPL) sentence — the textual rendering. The two modes are semantically equivalent per the standard's bimodality principle; V6 (OPL syntax conformance) and V7 (OPD–OPL coherence) validators enforce equivalence on every OPD-OPL pair in this chapter.
The standard's normative content used here:
- §6 Object-Process Diagram (OPD) — diagram syntax: object rectangles, process ellipses, structural and procedural links, in-zoom and out-zoom refinement.
- §7 Object-Process Language (OPL) — natural-language form: sentence productions R1–R20+ in Annex A's EBNF.
-
§8 Generic OPM ontology —
essencepredicate (physical | informatical),affiliationpredicate (environmental | systemic), structural relations (exhibits,consists of,is a), procedural relations (requires,yields,affects,handles,invokes).
The OPDs are organized as a progressive in-zoom hierarchy per §6.5 of the standard: the top-level System Diagram (SD) presents the system at its highest abstraction; subsequent diagrams (SD1, SD2, …) in-zoom into a specific top-level component or process. Each OPD's narrative below names which top-level entity it refines, cites the architectural commitment it upholds, and cross-references the arc42 chapter (sections 3–8) elaborating its content normatively.
The conceptual model is load-bearing for Prism's cost commitments — the architecture (chapters 4, 5, 8, 9) and the implementation must uphold the cost-model the OPM establishes here, not the reverse. Four cost-model commitments are normative at the conceptual layer:
For a PrismModel declaring a TypedCommitment carrying K bits of typed-bandwidth admission predicates per ADR-048, the expected number of forward(input) invocations to find a κ-label satisfying the commitment equals α⁻¹ × 2^K at the σ-Projection Hardening Principle's U6 bandwidth-additivity axiom (ADR-047), where α is the admission probability of the base relation σ(x) ≤ target declared via ConstraintRef::Bound per ADR-040. The relation is tight (equality, not upper bound) — the Lean tight-bound theorem Commitment.prf_prob_tight_wellFormed (foundation-side Lean corpus) is the formal proof.
The conceptual commitment: the host's retry loop count is the only externality; everything else is structural. The architecture's role (chapters 5, 9) is to monomorphize the typed commitment so its evaluation adds no per-invocation cost beyond the K inlined predicate evaluations; the implementation's role is to realize this monomorphization through the substrate's TypedCommitment trait surface (ADR-048).
A forward(input) invocation that produces Ok(Grounded<Output>) performs:
- One catamorphism traversal of the verb body's Term arena per ADR-029 (stack-only; bounded by the Term arena's depth).
- Eight resolver-bound ψ-stage emissions per ADR-035 (the canonical k-invariants branch: ψ_1, ψ_7, ψ_8, ψ_9, plus four off-path stages for substitution-axis completeness per ADR-036).
- K inlined
ObservablePredicate::evaluatecalls per ADR-049 inside the post-resolverC::evaluate(kappa_label)consultation per ADR-048.
No heap allocation along the principal data path (TC-01 + ADR-019 + #![no_std] posture). No dynamic dispatch (no Vec, no dyn per ADR-048's rejected alternative 2). No runtime-unknown loop bounds (every loop is bounded by a HostBounds associated constant per ADR-037 or a ConstrainedTypeShape::CYCLE_SIZE per ADR-032). Per-invocation cost is O(1) + K · t_eval where t_eval is the per-predicate observable-read cost (O(digest_byte_length), a single pass over the digest bytes).
Structural mechanism — the Fold-Fusion Principle (ADR-054) with universal axis-body enforcement (ADR-055). C2 is realized by fold fusion: every prism transformation — verb-body composition, prism_model! route body, resolver body, and every axis-impl body — is a folding operation, and the catamorphism's universal property (ADR-019) fuses composed fold-rules into a single structure-preserving map. Three fusion layers stack: (i) macro-expansion fusion — verb-call sites lower to Term subtrees at proc-macro expansion (ADR-024); after expansion the verb-call walker is absent at run time. (ii) catamorphism fusion — the per-variant fold-rules (ADR-029) compose universally; the initial-algebra property guarantees there is no choice about how to fuse them. (iii) monomorphization fusion — the resulting per-substitution-axis-selection Rust routine has no dynamic dispatch and no allocation (TC-01); LLVM inlining collapses per-variant dispatch with fold-rule bodies into straight-line code. Per ADR-055's universal substrate-Term verb body discipline, every AxisExtension impl (standard-library AND application-author custom) carries a substrate-Term verb body via the foundation-declared SubstrateTermBody sealed supertrait bound — the catamorphism's Term::AxisInvocation fold-rule recursively folds the axis's body_arena() static Term slice rather than calling an opaque kernel function. Fold-fusion's structural reach extends to the leaf level across every axis surface for every axis impl. There is no carve-out for "novel-domain axes": per ADR-019's initial-algebra commitment, every prism operation is substrate-Term decomposable; operations that cannot be decomposed are not prism operations and must be hosted at the host boundary outside the catamorphism's reach. C2's "one Rust function per forward invocation" claim is structurally guaranteed, not provisional — fold-fusion holds universally for every prism transformation, enforced at compile time through the SubstrateTermBody supertrait bound.
The conceptual model commits Prism to one structural inference per input at the typed-iso surface. The catamorphism's per-variant fold-rules per ADR-029 traverse the Term arena deterministically; the resolver-internal iterative-resolution discipline per ADR-043 is not a search over the typed-iso surface — it is the substrate's structural-transform realization machinery (the ψ-residuals discipline per ADR-035 + ADR-046 forbids search-shaped emissions in verb bodies; the resolver-internal loop is admissible because resolver bodies are not closure-body expressions per ADR-046).
The conceptual commitment: admission is a structural property of the value's k-invariant signature, not a runtime predicate over byte values. The host-boundary retry mechanism (varying the input outside the iterative-resolution domain) is external to the catamorphism's per-invocation contract.
The canonical hash axis selection (the third substitution position per ADR-030) must satisfy the σ-Projection Hardening Principle's six axioms (U1–U6 per ADR-047) for the cost-model commitments above to hold. A non-conforming Hasher impl introduces a cheaper-than-σ observable predicting admission, which invalidates C1's equality: the empirical trial count falls below α⁻¹ × 2^K (a structural attack exploits the non-conformance). U1–U6 are conceptual axioms on the projection's structural properties; the empirical witness through axis::cryptanalyze per ADR-049 audits the axioms; foundation's CI enforces conformance on every release.
The OPDs below proceed from the most-abstract (SD, system context) to specific in-zooms (principal data path, verification, compile-time enforcement, distribute-and-run). Each OPD's narrative names which cost-model commitment(s) (C1–C4) it upholds and cross-references the arc42 chapter elaborating its content. Read SD first; then read the subsequent SDs in the order they appear; the conceptual layering matches the architectural exposition.
Generated from sources at UOR-Framework.wiki. Do not edit pages directly via the GitHub web UI — edits are overwritten by the next build. See README for the authoring workflow.