Skip to content

ADR 032 Methodology Agnosticism as Foundational Axiom

Michael Zargham edited this page May 18, 2026 · 1 revision

ADR-032: Methodology Agnosticism as Foundational Axiom

Status: Accepted Date: 2026-05-18 Deciders: Michael Zargham Related: ADR-003 Topological Framework Documented as Future Work; ADR-020 Vocabulary Alignment with Zargham 2026; ADR-029 Engineering Lifecycle Stages as Scope Metadata; ADR-030 Polycentric ASOT Authority Model; ADR-031 Attestation Status Pass Fail Deferred Deprecated; Mission and Thesis; Topological Framework Future Work; Design Spec

Context

Several locked decisions in v0.1 — the polycentric ASOT model (ADR-030 Polycentric ASOT Authority Model), the optional lifecycle vocabulary (ADR-029 Engineering Lifecycle Stages as Scope Metadata), the attestation status that handles regression locally (ADR-031 Attestation Status Pass Fail Deferred Deprecated) — share a common underlying principle: flexo-rtm is methodology-agnostic. Different programs use different assurance methodologies (INCOSE / ISO 15288, DO-178C, NASA Phase A–F, ISO 9001, Agile, MIL-STD-498, customer-program-specific milestones, or no formal methodology at all), and the framework should support all of them on the same footing rather than privileging any one.

The methodology-agnosticism axiom has not been stated as a first-tier design principle in its own right. As a result, accumulated wiki drift has subtly positioned the topological framework articulated in Zargham (2026) as the planned future of flexo-rtm — "what v0.1 defers until the registry conversation happens." That framing overstates the relationship. The topological framework is a related research line with shared philosophical kinship (named-approver discipline, structural accountability, V&V distinction); it is not what flexo-rtm is for. By definition, any topological audit operates downstream on data flexo-rtm produces; the audit does not produce the data.

This ADR makes the axiom explicit, names what flexo-rtm IS, and clarifies the relationship to the topological framework so future contributions do not drift back into "topology is the eventual goal" framing.

Decision

Methodology agnosticism is a foundational design axiom for flexo-rtm. Concretely:

  1. What flexo-rtm IS. A working tool for bidirectional requirements traceability that is reduced to practice, plus a record of human signers where judgment is rendered. Both halves are settled: forward and backward traces with coverage statistics are decades of established RTM practice (Doors, Jama, Polarion, OSLC-RM); structural enforcement of named approvers via SHACL is settled engineering practice grounded in W3C Verifiable Credentials Data Integrity, SLSA in-toto attestations, Sigstore + Fulcio, git GPG/SSH commit signing, NIST SP 800-63, and W3C SHACL. flexo-rtm composes these into RDF + named graphs for the SysMLv2 use case. Nothing in this scope requires a new research result.

  2. What flexo-rtm does NOT commit to. Any specific assurance methodology beyond the bidirectional-traceability + named-signers floor. INCOSE / ISO 15288 stages are one example of an engineering lifecycle vocabulary (see ADR-029 Engineering Lifecycle Stages as Scope Metadata for the methodology-neutral framing); DO-178C, NASA Phase A–F, ISO 9001, Agile, MIL-STD-498, and custom phasing participate on equal footing. The topological framework is one example of an assurance methodology that some adopters may find useful; the framework is not required and the framework is not what flexo-rtm is for.

  3. The topological framework's relationship to flexo-rtm. Zargham (2026), Formalizing Document Assurance: A Topological Framework for Verification, Validation, and Human Accountability, articulates a typed simplicial complex framework for assurance that shares philosophical elements with v0.1's accountability discipline — both treat named-human approval as structurally load-bearing rather than recommendation-level. The framework is a separate, related line of inquiry, not yet reduced to practice (the registry-of-pre-approved-types question, the alternative-invariants question, and the recursive-completeness question are all open research). If the framework matures, it operates as downstream analysis on data flexo-rtm captures — taking the traceability and attestation graph as input and computing additional structural properties over it. flexo-rtm provides the substrate; the framework would consume the substrate. The two are not the same thing, and the substrate does not depend on the analysis.

  4. Vocabulary that aligns with the topological framework remains in v0.1 ontology (per ADR-020 Vocabulary Alignment with Zargham 2026) because the alignment is cheap, the philosophical kinship makes some shared terms useful (Guidance, AdequacyCriteria, SufficiencyCriteria, Aspect), and adopters who want to feed their data into a topological analysis later can do so without translation. This is forward compatibility for an optional analysis path, not "carrying forward the v0.1 vocabulary to its eventual destination."

  5. Other methodologies receive the same forward-compatibility treatment. An adopter wiring flexo-rtm to a SLSA supply-chain audit pipeline, a Goal Structuring Notation assurance case builder, an ARP4754A airborne-systems audit, or a custom in-house analysis layer benefits from the same substrate. flexo-rtm does not privilege one downstream-analysis target over another.

Consequences

Positive

  • The framework's relationship is honest. Adopters reading the wiki understand that flexo-rtm is a working tool that stands on settled practice, not a half-built prototype on the way to a topological audit. The topological work is one related research line with philosophical kinship, not the destination.
  • Methodology choice is the adopter's, not the framework's. Programs already running DO-178C, NASA Phase A–F, ISO 9001, Agile, MIL-STD-498, or custom phasing adopt flexo-rtm without translating their methodology into a privileged one. This is consistent with the polycentric ASOT model (ADR-030 Polycentric ASOT Authority Model) and the attestation-status mechanism (ADR-031 Attestation Status Pass Fail Deferred Deprecated).
  • The research line is unburdened. The topological framework progresses on its own honest timeline (the registry question, the invariant question, the recursive-completeness termination) without flexo-rtm adoption depending on its readiness, and without flexo-rtm's release schedule pressuring the research.
  • Vocabulary alignment is preserved. Guidance, AdequacyCriteria, SufficiencyCriteria, and Aspect remain in v0.1 ontology as forward-compatible alignment to the topological framework's vocabulary (per ADR-020 Vocabulary Alignment with Zargham 2026); adopters who later run topological analysis benefit from the alignment, and adopters who never do are unaffected.

Negative / Tradeoffs

  • Some readers may have expected flexo-rtm to be "the topological framework, eventually." The wiki has occasionally implied that framing; this ADR makes the actual scope explicit. Readers expecting an eventual topology-first product will need to recalibrate.
  • A lower-ambition framing. "Bidirectional traceability + named signers" is a deliberately modest scope. The user-facing tradeoff: a more ambitious scope claim would be more rhetorically impressive but would not be honest about what flexo-rtm actually delivers in v0.1 (or what it needs to deliver to be useful in practice).

Neutral

  • The existing locked decisions (ADRs 029, 030, 031) are all specific applications of this axiom. This ADR names the underlying principle they share; it does not change any of them.
  • The Topological Framework Future Work page remains in the wiki as the canonical reference for the related research line. Its framing is updated to match this ADR: the page documents the topological framework as a related line of inquiry, not as flexo-rtm's planned destination.

Alternatives Considered

  • Leave the axiom implicit. Rejected. The wiki has accumulated drift implying topology is the planned destination of flexo-rtm; without an explicit axiom, that drift recurs.
  • Commit flexo-rtm to the topological framework as the eventual audit primitive. Rejected. The framework has open research questions (registry governance, alternative invariants, recursive-completeness termination) that are not on flexo-rtm's critical path. Coupling the framework's research timeline to flexo-rtm's release timeline harms both.
  • Drop vocabulary alignment with Zargham 2026. Rejected. The alignment is cheap (a handful of class names), the philosophical kinship makes some shared terms useful (Guidance, AdequacyCriteria, SufficiencyCriteria), and adopters who later run topological analysis benefit from the forward compatibility. Dropping the alignment would pay no benefit and would lose the easy interop with that research line.
  • Commit to a different specific methodology (DO-178C, NASA, ISO 9001, etc.). Rejected for the same reason: privileging any one methodology shrinks the adopter set unnecessarily. Methodology-agnostic substrate + adopter-specific downstream analysis is the right cut.

Implementation Notes

This ADR is a framing decision rather than a code change. Its consequences land as wiki edits across many pages, removing language that positions the topological framework as flexo-rtm's eventual destination and replacing it with language that names it as one related line of inquiry. Specifically:

  • Mission and Thesis Proposition 6 reframed to lead with methodology agnosticism; the topological framework moves from "future work, not v0.1" to "one example of an assurance methodology flexo-rtm does not commit to."
  • Topological Framework Future Work opening reframed: the "original aspiration was to make the typed simplicial complex the primary audit primitive" statement is excised (it was a misinterpretation; that was never the aspiration). The page is repositioned as the canonical reference for the related research line.
  • Map of Content, Gap Taxonomy, ADR-005, ADR-017, ADR-020, ADR-021, ADCS Prototype Lessons, External URI References, Verifiable Self-Certification, and similar pages sweep-updated: "future framework" → "topology line of inquiry / downstream analysis"; "when the framework lands" → "if the framework matures, as one downstream analysis path."
  • v0.1 ontology continues to carry the Guidance / AdequacyCriteria / SufficiencyCriteria / Aspect / AssuranceComplex / AssuranceFace / AssuranceTriple vocabulary, framed as forward-compatible alignment to the topological framework should adopters choose to run downstream analysis under it.

References

Start here

Foundation

Internal Research

External Research

v0.1 Certification Model

Roadmap and future work

Three-Layer Architecture

Ontology Design

Interface Contracts (normative)

Adapter Contracts (background)

Reproducibility

Open issues

Decision Log

Meta

Clone this wiki locally