Told, shown or measured, and how current each one is.

Every claim in NovusLens carries a rungRungOne of three levels of support for a recorded value: told, shown or measured. The level appears beside the value. on the evidence ladderEvidence ladderThe three levels of support a record can have, in order: told, shown and measured.. The rung says what kind of support the claim has. It does not say that the claim is correct.

• toldToldEvidence that somebody stated the value. It is the weakest level, and it is always attributed to its speaker. means someone stated it. An interview answer, or a statement made in a meeting, is told. It is attributed to its speaker.
• shownShownEvidence that a document supports a value. A recovery plan that states a target shows the target, but it does not show that the target was achieved. means a document supports it. A recovery plan that states a one-hour target is shown. The plan does not show that the target was achieved.
• measuredMeasuredThe top evidence level: a connected source recorded the observation within a stated period. A measurement shows what was observed on that occasion, not that it always holds. means a connected source recorded an observation within a stated period. An incident record that shows a recovery took two hours on one occasion is measured for that occasion.

The product uses these three words in this sense only. Read the rung before you read the claim.

An example

Three records might sit under one claim about recovery. A leader says the team recovers within an hour. That is told. The recovery plan gives an hour as the target. That is shown, and it leaves open whether the target was met. The incident log records two hours on the most recent failure. That is measured, for that failure, within the period the log covers.

The three do not overwrite one another. All three stay on the record, each with its rung, source and date. The rung tells you which one to rely on for which question.

How to read it

Check three things before you rely on a claim: the source and its date, what the source covers, and whether another record disagrees with it.

The rung also limits what can be built on a claim. The measurement calculations use only the rungs they are allowed to use. For example, a Complexity, Agility and Recoverability (C/A/RC/A/RThree separate scores for a team or service: Complexity, Agility and Recoverability. Each uses current evidence and the customer’s stated scoring rules.) score needs a calibrationCalibrationThe customer's own rule that turns raw figures, such as a gauge, a workflow history or an incident log, into a C/A/R score. A score moves when the calibration moves, and the change is recorded. supported by a document. If required evidence is only told, the score is withheld. A critical path length (CPLCPLCritical path length. The total coordination weight of the longest chain of dependencies in the model. NovusLens reports it as index points, not elapsed time.) figure uses a measured link only where the measurement is current, and otherwise keeps the told account of that link.

How long a measurement stays current

A measurement has a window. By default an observation from a connected source is current for 30 days after the source wrote it. Setup’s advanced settings change the length, which .novuslens.env saves as CONWAY_OBSERVATION_TTL_DAYS.

When the window closes, the observation is not deleted and it is not quietly kept in the current picture. It stays in the history as measured, is excluded from the current view, and is counted as expired. The stored critical pathCritical pathThe dependency chain with the greatest total coordination weight in the map. It identifies a structural constraint. It does not state elapsed time. record counts the expired observations it left out.

A feed from a supplier outside the organisation uses its own refresh interval, which the source declares. A publication older than three of those intervals is refused when it is read, so a stale feed cannot keep supporting a findingFindingA difference in the records that needs a decision. It includes its sources and can describe a slow handoff, disputed fact or missing owner.. The evidence-state guide explains how those findings are withheld when their support runs out. The operator should read the feed on the agreed schedule.

Conflicts are kept, not overwritten

When two records disagree, NovusLens keeps both. If two versions of a record differ, the C/A/R calculation keeps both and records a conflicting-evidence finding in its results. Neither version enters the calculation until the conflict is resolved. A disputed link in the critical path calculation is counted at its heavier weight, so the figure does not quietly take the lighter account.

The mapMapThe view of teams, services and their dependencies for the selected part of the organisation. Details appear beside the map. keeps different accounts of the same fact side by side. When a person records a resolution to a challengeChallengeA recorded dispute of a value. Both the original and the disputed version are kept, and the challenge is listed under Findings., both accounts stay in its history. A newer statement does not silently replace a value supported by a document or a measurement. The older value remains available, and corrections or withdrawals change what currently supports a conclusion.

Simulated observations

Some records are labelled as simulated. They come from a demonstration source. They exercise the real software path from intake to findings, and they do not describe your organisation’s performance. The label is carried on the record. A feed that carries the simulation marker is refused by a production enrolmentEnrolmentThe operator's recorded permission for a source to send specified record types about named agents, using an identified signing key., so simulated data cannot be read as customer measurement.

Citations resolve to exact bytes

When a claim cites a document, the citation is tied to the stored bytes by a content addressContent addressAn identifier calculated from a record's contents. It lets the program check that the retrieved bytes are the same bytes originally stored.. A citation either resolves to the exact stored content or fails with an error. A purgedPurgeRemoving a stored document’s contents while keeping a record that it was removed and why. Conclusions that used it must be checked again. document keeps a marker that it was removed, so what it supported does not come back by accident. If the same bytes are acquired again after a purge, the copy gets its own identifier.

External suppliers

The same ladder applies to suppliers outside the organisation. A supplier’s stated service level is a declared claim. A figure measured by the organisation’s own systems is the observed view. NovusLens compares the two. If the measured figure falls below the declared target, it raises a finding for review. This comparison applies to agents and other suppliers.

The first step, a black box, means NovusLens has no declared or measured view of the supplier. For an agent, intake checks its self-declared document. Once verified, the document moves the recorded view from black box to declared. It is recorded at the shown rung and cited.

Limits

A rung describes support, not truth. A measurement can be measured and still be the wrong thing to measure. A calibration written by the organisation makes its assumptions visible, and it does not make them right. The product shows the rung and the window. It does not remove the need to check the source.

What to do next

Open the source behind any claim you plan to use. Check its date and the period it covers. If a value looks wrong, ask the source owner to review it and to publish a correction or withdrawal. Then check the affected figures again, because the rung of a claim can change when its support changes. For how this fits the wider stance on measurement, read how NovusLens measures and what it does not claim.

Where it is implemented

• apps/graph/src/conway_graph/evidence.py stores evidence by its content address, keeps a removal marker for purged documents, and returns the cited content or an error.
• apps/reasoning/src/conway_reasoning/evidence_view.py maps the internal labels to the words told, shown and measured (RUNG_WORD) and shows each source only in the form it can safely take.
• apps/graph/src/conway_graph/engine.py maps the stored rungs to words, counts measured links on a path, and excludes expired observations from the current picture.
• apps/connectors/src/conway_connectors/corrector.py sets the 30-day default window (OBSERVATION_TTL_DAYS).
• apps/connectors/src/conway_connectors/boundary_contract.py refuses stale publications and simulated records in production enrolment.
• apps/graph/src/conway_graph/car_assessment.py keeps conflicting versions and raises conflicting-evidence.
• apps/graph/src/conway_graph/governance.py compares declared service levels with boundary-measured figures for external nodes.