RockWorx Acquisition Ontology + SysML v2 Link

A BFO/CCO-conformant ontology of the U.S. DoD acquisition lifecycle -- plus a legislative-horizon layer that models what Congress is about to change, and a working, standards-based link that resolves SysML v2 system models into it. Published as three ontology modules plus two reproducible demonstrators. License: BSD-3-Clause. Creator: RockWorx Aerospace.

• Base: https://w3id.org/rockworx/acq (rwx-acq-base.ttl) -- what IS law/policy.
• Transform: https://w3id.org/rockworx/acq/transform (rwx-acq-transform.ttl) -- the 2025 reform.
• Legislative horizon: https://w3id.org/rockworx/acq/horizon (rwx-acq-horizon.ttl) -- what MIGHT become law: pending bills as non-asserted candidate changes. Demonstrator + pipeline in horizon/.
• SysML v2 link + demonstrator: sysml_link/

Why this exists

The Department of Defense has directed BFO + CCO as its baseline for formal ontology and is moving to SysML v2 for digital engineering -- yet no public ontology models the DoD acquisition lifecycle on that base, and nothing links a SysML v2 system model to acquisition-process semantics. This fills both gaps:

• The ontology -- what the acquisition process is: the DoDI 5000.85 Major Capability Acquisition lifecycle (phases, milestone gates, technical reviews, program risk, funding, oversight) plus the 2025 Warfighting Acquisition System / Joint Force Requirements Process reform, grounded to CCO 2.1 + BFO 2020.
• The link -- so a program's actual SysML v2 model becomes queryable against phases, gates, reviews, KPPs, program risk, and Title 10 authority.

The ontology headline: the ICE -> Objective shift

The most important thing to understand before querying is that the post-2025 reform did not just rename a document -- it changed the KIND of thing the requirements root is. Under JCIDS the Initial Capabilities Document is a Prescriptive Information Content Entity (CCO ont00000965) -- it implies a materiel solution. Under the Joint Force Requirements Process the root is an Objective (CCO ont00000476) -- the Key Operational Problem, which states the problem and withholds the solution. These are different OWL/BFO classes; the transform module marks the legacy term owl:deprecated with an IAO "term replaced by" pointer, but term-replaced-by is not sameness of kind. Query across the reform with the role-stable isRequirementsInputTo property, not by class name (see RELEASE_NOTES.md and bridge-queries.rq). Full module detail, conformance notes, and provenance: RELEASE_NOTES.md.

SysML v2 integration (sysml_link/)

SysML v2 gives a standard, ISO-track, vendor-neutral system model behind a standard API -- but a model is only meaning-bearing if its elements resolve to shared, directed semantics. This link supplies exactly that, using SysML v2's native constructs:

• The RockWorx acquisition classes are emitted as a native SysML v2 tag library (owl:Class -> metadata def, owl:ObjectProperty -> connection def).
• A model tags its elements with those classes and uses first-class SysML v2 constructs -- a requirement (the KPP), satisfy (design satisfies requirement), and a verification case that verifys the requirement with a VerdictKind result.
• A transform reads the model's AST and resolves each tagged element and native relation into the acquisition ontology as RDF. Acquisition-process relations that SysML v2 has no native construct for (a requirement assessed at a review, a review that informs a gate, a KPP that traces to Title 10 authority) are carried by a small bridge overlay. That division -- native SysML v2 for the system model, the bridge for the acquisition process -- is the point.

The demonstrator ("Risk-Driven Milestone Validation"). A tiny SysML v2 model whose KPP verification records a FAIL. From that failure a program risk is derived (not hand-asserted), and a SHACL rule enforces the acquisition decision: a single SPARQL query answers "what is preventing a successful Milestone B decision?" end-to-end (KPP -> review -> gate + failed verification -> derived risk + Title 10 authority), and the Milestone B gate is blocked while the risk is open -- and passes once the design is fixed (result -> PASS) or the risk is formally accepted. Reproduce it Python-only (rdflib + pySHACL) from the committed artifacts; see sysml_link/ and its reference write-up.

Warranted scope. This is a buildable, standards-based, DoD-directed link demonstrated on one small model -- not a finished platform. It rides the standard SysML v2 API (any conformant tool, no lock-in), uses native v2 metadata (no brittle vendor mapping), and lands on the DoD-directed BFO/CCO base.

Legislative-horizon layer (horizon/)

Base + transform assert what IS law/policy. The horizon module (rwx-acq-horizon) adds a THIRD, prospective layer: what MIGHT become law. Pending U.S. defense-acquisition bills are modeled as non-asserted candidate changes, so a program's acquisition context can be read against not just today's rules but the rules Congress is poised to change.

• The credibility linchpin -- prospective != asserted, machine-proven. OWL is monotonic and has no modality, so the layer keeps prospective strictly out of the asserted model: candidate changes are Descriptive ICEs owl:disjointWith the base roles; the affected construct is referenced by OWL 2 punning (no propagating semantics); maturity/likelihood/impact are annotation properties (a judgment can never be entailed as a fact); graduation on enactment is a governed append-only migration, not an inference. Reasoning base + transform + horizon + the demo A-Box yields no new class subsumption among base/transform classes (verified in horizon/qc/).
• The demonstrator ingests real bills from the Congress.gov API (POLL real status -> SHRED provisions with an LLM, grounded in the text -> TAG maturity + a sourced likelihood judgment -> LOAD non-asserted). It answers "what is on the legislative horizon for Milestone B / the MCA spine, and how likely?" and shows the FY26 NDAA (PL 119-60) codification of the Portfolio Acquisition Executive graduating from policy (EO 14265) to statute. Reproduce Python-only: python -m horizon.run_demo. See horizon/README.md.

Provenance

• Grounded to CCO 2.1 + BFO 2020 (via owl:imports), to DoDI 5000.85 for the MCA spine, and to the 2025 reform's directing instrument, Executive Order 14265, "Modernizing Defense Acquisitions and Spurring Innovation in the Defense Industrial Base" (9 April 2025).
• Conformance-reviewed. Both modules and the SysML v2 link went through RockWorxOS-authored BFO/CCO conformance passes under the RockWorxDuo review protocol -- design-level and output-level cross-checks, plus an adversarial red-team pass at a public-release bar. These are conformance reviews, distinct from the machine-checked tooling QC below.
• Machine-checked. Both modules reason ELK-consistent, zero unsatisfiable classes, and the demonstrator's instance graph reasons consistent against the full BFO/CCO + module stack (qc/, fully offline-reproducible). ROBOT reports clean apart from a deliberate, documented divergence around the retained deprecated-class axioms (see RELEASE_NOTES.md).

Note on department naming: the statutory name is the Department of Defense (used by EO 14265). Executive Order 14347 (5 Sep 2025) authorizes "Department of War" / "Secretary of War" as an additional secondary title for public and non-statutory contexts; the two co-exist. This ontology quotes each instrument in its own language.

License

BSD-3-Clause. See LICENSE. Copyright (c) 2026 RockWorx Aerospace.

Resolving the IRIs

https://w3id.org/rockworx/acq* are persistent identifiers under the w3id.org service; the redirect is registered by a pull request to perma-id/w3id.org at publication time. Until it merges the IRIs are valid, stable identifiers that do not yet resolve over HTTP -- use them as-is for class/property identity and owl:imports.