Silence is a bound

In our published worked service, one supplier’s silence is worth 21.5 reach points. The dominant lineage’s upstream reports an evidenced reach of 57.1% and a bound of 78.6%. The distance between the two figures has a single cause, a component whose ancestry nobody has disclosed. The service’s two-hour impact tolerance passes on the evidenced domain and fails on the bound, so the pass-fail outcome of a regulatory test depends on that one gap.

A procurement team can put that number in front of a supplier.

One rule, applied everywhere it binds

The framework states one evidential principle and applies it everywhere it binds. An unknown is never recorded as a favourable value.

The principle appears three times, in three different mechanisms, because it binds in three different places. A value with no stated evidence tier is treated as inferred, the lowest grade, since untiered data is inferred data whose inference was not written down.

A service with no defined target state reports its coverage as not computable rather than computed against an implicit definition. And a component whose lineage is undisclosed stays in the index as itself while every reach bound that cannot exclude it carries it, which is where the 21.5 points come from.

The reviewer error catalogue records what happens when the principle slips. Its last row is a layer-6 cell showing a low index with no entropy data, read as diversified. Absence recorded as the best possible news inverts the finding entirely, and the assessor who wrote it down never notices, because nothing on the page looks wrong.

What silence does not do

Non-disclosure never moves the index, and the reason is structural rather than generous. The undisclosed unit stays in the computation as itself, because the index measures the estate and the estate did not change when the supplier declined to answer. What widens is the bound, and the spread between known reach and the bound is the measurable cost of the non-disclosure, denominated in reach points. A procurement conversation can start from that number alone.

The framework’s determination log holds a worked example of the boundary case. A supplier answering that it uses “industry-standard, FIPS-validated cryptographic implementations” is recorded as undisclosed with status declined, because a validation confirms correct computation and says nothing about ancestry. Accepting marketing language as disclosure has its own row in the error catalogue, and the row exists because the answer sounds so reasonable that assessors accept it every week.

Auditing reached this position long before we did. Under the international auditing standards, an auditor who cannot obtain sufficient appropriate evidence does not issue a clean opinion and move on. The scope limitation is stated, the opinion is qualified or disclaimed, and the missing evidence is never resolved in the client’s favour.

A second line asking suppliers for lineage evidence is running the same discipline one level down, and the reach bound is its qualified opinion, stated in the service’s own units.

The silent share is a number

The principle scales past a single assessment. Our census methodology enumerates validated payment modules from public certification listings and grades every lineage claim, and the register of modules whose security policies name no upstream publishes with each snapshot. The silent share is itself a published figure, because the absence of a named upstream is never evidence of an independent implementation.

The Supplier Lineage Disclosure Request gives every supplier three ways to appear in that record. Attributed disclosure is preferred, and it is not the only route. A supplier declining attribution may elect anonymised aggregate publication, entering the figures without a named row. A supplier taking neither option is recorded as a named non-response, with a factual-review extract provided at least 14 days before anything publishes. They were asked, and the record shows what they said.

Silence stays the one answer that costs a supplier nothing to give, so the framework makes it the one answer with a published price. Disclosure converts an undisclosed entry into a computable one, moves a bound, and never worsens an index, which is why lineage disclosure belongs in procurement and due diligence rather than in audit season.

The two figures were designed so that answering is always the better trade.