SPAN
← Back to GT Labs

SPAN IN ACTION · RECORDED MODEL EXECUTION

When the plan changes,
control stays with the work.

One research request. Independent providers. A failed testing step.
Watch the work adapt, obtain revised approval, and reach a verified outcome.

RNA-001 / RESEARCH WORKFLOWReady
RESEARCH TEAMPrepare, test & assessOutcome + satisfaction conditions
SPANCoordinates work · checks authority · evaluates evidence
A

Preparation

Capability advertised
B

Testing

Capability advertised
D

Alternative

Capability advertised
C

Assessment

Capability advertised
○ Preparation accepted○ Testing accepted○ Assessment accepted
1 / 30

WHAT TO WATCH FOR

The difference is
in the decisions.

  1. Completed work is retained. The preparation remains valid after the testing provider fails.
  2. A new path needs new authority. The alternative costs more and has a later testing window. Execution waits for the modelled approval.
  3. Evidence determines completion. The request is satisfied only after the required results have been accepted.
Inspect the recorded state and decision trace +

These snapshots were generated by executing the SPAN Visualiser v0.6.1 recovery scenario. Playback follows that recorded run; it does not execute a live distributed network.



DEMONSTRATION SCOPE
Real coordination code. Modelled participants.
The Python model generated these state transitions. The providers, identities, trust checks, and receipts are simulated. This is an illustrative research workflow, not a biomedical simulation or a live customer deployment. Distributed enforcement, cryptography, live-system integration, and independent security evaluation remain engineering milestones.