When the digital baseline meets the lift, fixture, and flight line

This week, the Air Force finished vertically stacking an inert Sentinel missile after a four-month campaign spanning transport, field assembly, interface checks, subsystem integration, and design review. The important signal is not that a digital model became hardware. It is that the program had to prove the technical baseline, logistics pipeline, handling process, people, and assembled configuration together before flight testing.

That is the DDNA thesis in physical form: a model is useful only when requirements, design, material, process, tooling, test evidence, and configuration remain coherent as the product moves from concept to scale.

Inside DDNA#

Bring a repository into DDNA without silently starting work#

Connected Gitea repositories can now enter DDNA as restricted, explicit-start projects. The project surface shows the source repository, lets an authorized user stage supporting documents, and keeps generation behind a deliberate Start action. Repository content is tied to immutable commit and blob identities so the source state used by downstream engineering work can be reconstructed later.

The intake interface is visible in production. This week’s review verified the ready-to-start state and its repository, document, and Start controls; it did not execute a customer project or claim that every repository has completed generation.

Update the affected Intent record—not the entire program narrative#

New evidence can now trigger a focused Intent update. Draft material can be updated directly, while approved or in-review material receives tracked suggestions for human review. The workflow preserves unchanged blocks, tables, diagrams, source identities, and revision history instead of treating every new upload as a reason to regenerate the whole document.

This is a deeper change than faster document creation. It turns evidence arrival into controlled engineering change: identify what the source affects, propose the smallest coherent update, preserve the rest of the baseline, and leave an auditable decision trail.

Ask for a document change and review the diff#

Project Intent chat can now draft document edits as reviewable changes. When a request spans more work than one execution slice, the workflow can continue automatically while retaining project context. The result is still a proposal: people review each diff before it becomes part of the controlled record.

Generate alternatives, then make the decision explicit#

Market Intelligence runs can produce three visual approaches for the same evidence set. DDNA keeps the alternatives together, exposes the sources used for generation, and requires an explicit selection or refinement step before a visual is featured.

Enlarge

Enlarge

The comparison workflow is live in production. A selected visual is a presentation decision—not engineering approval—and inserting it into a document remains a separate controlled edit.

Engineering Radar#

Sentinel crosses from model maturity to field integration#

The Air Force completed vertical integration of an inert LGM-35A Sentinel at Vandenberg Space Force Base after a four-month campaign covering missile stages, propulsion elements, the shroud, transport, crane lifts, precision interface checks, and field assembly. A parallel flight-test design review assessed design maturity and hardware testing against performance, safety, and mission-assurance requirements ahead of planned 2027 flight testing.

For DDNA, the signal is the configuration boundary. The relevant product record is not only a missile design. It is the exact hardware, handling process, logistics state, personnel qualification, interface evidence, and review decision that made that physical configuration acceptable for the next gate. U.S. Air Force, Sept. 14

A B-52 modernization program has to change several aircraft at once#

The B-52 Commercial Engine Replacement Program passed Milestone B while its radar modernization passed Milestone C. The engine effort reaches beyond the engines into nacelles, inlets, hydraulics, displays, flight systems, generators, electrical power, and fleet structural assessments. Early wind-tunnel work exposed inlet distortion, forcing a redesigned nacelle and inlet before critical design review. Two test aircraft are receiving engine and radar modifications concurrently.

This is exactly where a document-centric digital thread breaks down. Each aircraft, modification kit, software build, power load, structural finding, and test result changes the evidence that is valid for the next decision. The engineering advantage comes from knowing which baseline a result proves—and which configurations it does not. U.S. Air Force, Sept. 15

Manufacturing preparation moves before machine delivery#

Siemens announced a “Meet at the Machine” initiative intended to let part manufacturers begin production preparation before a CNC machine arrives. The vendor’s claim is directional, not yet independent proof of productivity. But the architecture matters: part definition, machine capability, process setup, and production readiness are being connected earlier.

If preparation moves upstream, configuration control has to move with it. The machine that arrives must match the capability model used to prepare the job, and the resulting process plan must remain linked to the released part definition, tooling, inspection plan, and acceptance evidence. Siemens, Sept. 15

What to Watch#

1. SysML v1-to-v2 migration becomes an engineering-change program#

The Department’s current transition material frames SysML v2 adoption as a sequence: establish a transition team, develop a plan, conduct pilots, update infrastructure, and deploy to projects. That is more useful than treating migration as a file conversion. Requirements, element identities, interfaces, verification links, viewpoints, and approved baselines must survive the cutover—or be dispositioned explicitly.

Watch for: conversion acceptance criteria; persistent element identity across versions; traceability reconciliation; dual-baseline governance; and pilot results that test change propagation rather than diagram appearance. A counter-signal is a one-way export that produces attractive models but cannot demonstrate semantic equivalence or reconnect evidence. The Department’s page also keeps VV&A guidance, DoDI 5000.61/5000.66, and transition reports in the same digital-engineering context. DoD Digital Engineering, Modeling and Simulation

2. Legacy technical data can generate models—the hard problem is proving the model#

An open NAVSEA SBIR topic describes AEGIS documentation distributed across paper and legacy digital formats and seeks automated ingestion of drawings, installation and interface configurations, hardware and software bills of material, cable runs, IP data, protocols, and services into SysML. The desired outputs would support design, cyber assessment, interface verification, functional test, configuration management, reliability, and certification.

The requirement reveals the next bottleneck: model generation is not model authority. Watch for: source-to-element provenance; measured extraction accuracy by artifact type; treatment of contradictory revisions; human approval of inferred relationships; configuration-specific cyber and test views; and repeatable change impact when a source changes. The solicitation itself acknowledges that finished models may still require human modification. SBIR topic DON26BX05-NP003

3. V&V environments become controlled product configurations#

Autonomous and software-defined physical systems increasingly move through simulation, hardware-in-the-loop, range testing, and operational evaluation. Evidence only transfers between those stages when the system-under-test configuration, environment model, hardware, software, instrumentation, test procedure, and acceptance criteria are identified together.

Watch for: configuration IDs on every result; explicit model-purpose and validity bounds; recorded discrepancies between simulation, HIL, and live behavior; accreditation tied to a defined use; and review decisions that say where evidence is applicable. DoDI 5000.61, DoDI 5000.66, MIL-STD-3022, and the Department’s VV&A practices are strategic context here—not new standards, and not automatic certification. DoD engineering policy and guidance

4. Design for manufacturing moves into weapon-system architecture#

DARPA’s planned Next Generation Hypersonic Cruise Missile program is expected to pursue a fully integrated demonstrator while applying Design for Manufacturing and Assembly from program inception to establish a credible path toward affordable, high-rate production. The Industry Day is a pre-solicitation signal, not a completed program result.

Watch for: manufacturability requirements allocated alongside performance; material and process constraints represented before detailed design; tooling and supplier evidence connected to architecture decisions; production-rate assumptions tested with real process data; and verification plans that cover both mission performance and build repeatability. For DDNA, this is the shift from a digital model of the article to the product DNA required to reproduce it. DARPA NGHCM Industry Day

Wild Card#

Ground-qualified hardware is not space-qualified evidence#

NASA activated the Roman Space Telescope’s 300-megapixel Wide Field Instrument only after a staged in-space sequence: ten days of drying and decontamination, detector cooldown, calibration-system activation, zero-gravity operation of the element wheel, focus-mechanism checks, baseline starlight imaging, and thermal-control verification for the coronagraph. NASA reports the assessments are working as expected, with additional calibration still ahead.

The lesson is broader than space systems. Acceptance evidence has an environment and configuration. A ground test can support the next decision, but it cannot silently stand in for zero-gravity, thermal, integrated, or operational behavior. The reproducibility packet needs to say not only what passed, but under which physical state and use boundary. NASA, Sept. 15