Independent qualification means the outside organization controls something that matters
BTG-controlled tests, language ports and benchmarks are engineering evidence. They become independent evidence only when an unrelated organization controls a meaningful part of the methodology, environment, implementation, review or conclusion.
Five qualification modes
Independent implementation
Implement protocol behavior from published specification and conformance material without using BTG implementation source as the design basis.
Independent security review
Threat-model, attack, reproduce and report security findings under the reviewer’s own methodology.
Independent research benchmark
Select real workloads and comparison systems independently, then publish conditions under which ENTITY/BTDU helps or does not help.
Standards interoperability
Map ENTITY to an external standard, implement the interface and document semantic overlap, gaps and incompatibilities.
Institutional pilot
Apply a bounded ENTITY capability to a real organizational workflow with pre-declared success/failure criteria.
Evidence ladder
| Level | Evidence | What it supports |
|---|---|---|
| 1 | BTG deterministic tests and sealed evidence | Internal engineering behavior |
| 2 | External reproduction of public test campaign | Reproducibility in an unrelated environment |
| 3 | Unrelated implementation or integration | Specification/interoperability quality |
| 4 | Independent security/research assessment | Externally controlled critique and measurements |
| 5 | Institutional production-like pilot | Usefulness under real governance and operational constraints |
Publication discipline
- Participation does not equal endorsement.
- Negative findings remain publishable unless a pre-agreed responsible-disclosure window applies.
- Benchmarks must name workload, data shape, comparison architecture, hardware and measurement method.
- Legal ownership, regulatory compliance and external settlement are not inferred from software state alone.
- BTG-controlled reproducibility baselines remain labeled BTG-controlled even when technically strong.
Institutional research package
A university or research partner can begin with the Independent BTDU Qualification Program, then expand into authority, recovery, interoperability, data rights or agent governance. The preferred output is a citable report that clearly separates reproduced results, non-reproduced results, limitations and new findings.
Commercial pilot package
A commercial partner can begin with one operational boundary—credential-to-authority, agent authorization, data-rights exchange, provenance transfer, recovery or governed economic occurrence—and test it without adopting the entire platform.