The two original apps asked related review questions with different constructions. Health Atlas separated capability, maturity and performance gaps across six projects. The PMO put benefits, parties, evidence and interfaces alongside one another so a reader could try a scenario and ask what deserved further investigation. This Library home keeps both examples independent.

A picture that invites a question

The Health Atlas offered a visual fingerprint: three gap bars, a deforming tile, a tiny network and a sparkline. The deformation is an inspectable function of the supplied gap values. The network and trend were decorative examples, not the output of a dependency or measurement model. Their useful role is to invite the next question: what evidence, mechanism or history would this picture need before someone relied on it?

The three gap definitions remain distinct. Capability concerns resources, skills and tools. Maturity concerns how work is governed and controlled. Performance concerns delivered results. Their causes and possible responses cannot be inferred from three numbers alone. The original claim that leaders can spot outliers “in seconds” was an untested design aspiration.

A record is not the evidence it names

The historical Health Atlas roadmap proposed a “proof bundle” with three useful ingredients:

  • Acceptance artifact: a retained result of a specified check, together with what that check did and did not establish.
  • Rule receipt: a record of declared rules, inputs, versions and results, including conditions that were not met.
  • Lineage map: a trace from identified inputs through transformations to outputs.

Those ideas do not make a bundle proof of a decision’s quality. A hash identifies bytes; a recorded check concerns its stated criterion; a signature concerns the signer and signed content. None substitutes for an appropriate assessment method, sufficient project evidence, decision authority or evidence of benefit in use.

As an external conceptual reference, W3C’s PROV overview describes provenance through the entities, activities and people involved in producing something, and identifies interoperable representations of that information. This app does not implement PROV. Its versioned snapshots preserve toy inputs and recomputable results, not verified documents, an immutable ledger, a signature or a provenance graph.

The roadmap also favoured keeping the visual front door simple and detailed explanations inspectable on demand. Those choices are retained. Its proposed rule engines, automatic evidence collection, signed bundles, governance roll-ups, exports to enterprise systems and phased delivery commitments were never implemented in either source app. They are not imposed as a new research or product roadmap. The original specification remains recoverable at the pinned Git link.

Eight source lenses, without an integration claim

Original inspiration Useful idea preserved here What this PMO does not implement
Project Health Atlas A separate six-project gap picture A mapping from benefit scores to project gaps
Regulatory negotiation rehearsal Levers, parties and thresholds An agreement process, game solution or marginal pivotality
Governance trio Revisit an earlier calculation Independent evidence validation, rule engine or external lineage graph
Scenario / barrier / proof lattice Keep the three descriptions visible Formal concept analysis or computed barrier sufficiency
Portfolio Loom Compare supplied scenarios Temporal fibres, schedules or time-dependent resilience
Counterfactual programme steering Try a declared change An identified causal effect or structural causal model
Intent field navigator Ask how benefits relate to intentions A live CSV intent-mapping pipeline
Interface maturity simulator Ask about interfaces separately Coupled propagation between interface states

The original PMO README and source-app map used broader language such as integration, instantiation and spillover. The live source-lens view provides current links and the narrower, implemented relationship. The full original wording is recoverable from the source identities in the README. The three original seed files remain exact downloads; they were never runtime inputs to the source page.

What this example can and cannot answer

It can show the numerical consequences of supplied gaps and authored weights, preserve a state, replay its inputs and expose its arithmetic. It can make missing evidence or a claim’s boundary easier to discuss. No practitioner study, predictive validation, observed organisational agreement, actual benefit delivery or regulatory assessment has been supplied. The appropriate next move in a real review would depend on the specific question and evidence, not on the apparent precision of these scores.