docs(project): recount the release gates and re-target Track D at Flexo layer 1

[devin-ai-integration]

Summary

Two docs-only changes, both release-facing.

Recount the release-gate surfaces on main (README, spec-compliance.md, roadmap.md, training-examples.md), which had drifted well below what main actually carries: conformance 297→338, runtime robustness 165→194, golden ASTs 86→90, negatives 129→135, traces 98→106, tests/subtests 4,447→5,477. Measured from a green go test -race ./... plus the corpus (100/100) and stdlib (95/95) gates.

Correct the Flexo interop claim and rewrite Track D around the real target. The docs said matching Namespaces.kt meant "a graph produced here is addressed the way that service addresses elements". Reading flexo-mms-sysmlv2 and flexo-mms-layer1-service sources, that is false — namespaces match, addressing does not:

• extractModelElementToJson derives @id as substringAfterLast(':'), so elmt:Demo::Vehicle reads back as Vehicle and A::Widget/B::Widget collapse; requireValidId ([a-zA-Z0-9_-]+) rejects :: on every by-id path.
• Predicates outside sysml: and urn:sysmlv2:annotation:json: are ignored, so every sysx: triple (sourceText, memberIndex, hasBody, declaredKeyword) is dropped through that service.
• Paged listing/query key on sysml:elementId, which we never emit; roots filtering keys on sysml:owner/sysml:owningRelatedElement, so with only sysml:owningNamespace every element reads as a root.
• Multi-valued sysml: predicates are skipped in favour of the JSON annotation literal.
• Layer 1 itself constrains nothing about model vocabulary (ModelLoad.kt loads, diffs, commits; sanitizeCrudObject covers LDP CRUD objects only), so the contract lives entirely in the SysML v2 service's reader.

Track D is retargeted accordingly: the goal is stated as OpenSysML's Turtle standing in for flexo-mms-sysmlv2's output on a layer 1 branch graph, D3 is decomposed into D3.1 identity (name-based UUIDs, identity read from sysml:qualifiedName rather than the IRI) → D3.2 sysml:elementId → D3.3 ownership → D3.4 array/annotation duality → D3.5 the Fuseki+layer1 harness, a new D7 covers reference-valued properties emitted as strings and the abstract sysml:Import metaclass, and D1/D2 are reframed as interop prerequisites rather than nice-to-haves (since sysx: does not survive). Sequencing updated; stale R1/baseline text refreshed for 0.1.1 now that v0.1.0 is tagged upstream.

No behavior change; check-doc-links.py clean, no Go files touched.

Link to Devin session: https://nasa-jpl-demo.devinenterprise.com/sessions/6eb85b61128b4de8bdcfd7991202ccdc
Requested by: @HuiJun

[devin-ai-integration]

Original prompt from jason.han

@JPL-Devin/OpenSysML Help me adjust the 0.1.1 release that has been prepared

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR that start with 'DevinAI' or '@devin'.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment, CI, and merge conflict monitoring

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no potential bugs to report.

View in Devin Review to see 1 additional finding.