ArchSig Reference

Examples show how to read reports without overclaiming.

The example set uses a coupon feature extension to show split extension, hidden interaction witness, semantic witness, Feature Extension Report reading, and architecture dynamics interpretation.

Good extension

A split coupon extension interacts through declared interfaces rather than reaching into adapter internals. AIR can separate embedding, feature view, interaction, static split, and coverage claims.

Good extension shape
before:
  OrderService -> PaymentPort
  PaymentAdapter implements PaymentPort

feature:
  Coupon calculation

after:
  OrderService -> CouponPort
  CouponService -> CouponPort
  CouponService -> PaymentPort

Hidden interaction

A hidden interaction witness appears when the coupon feature reaches through a boundary, such as direct access to an internal payment cache. This is a static split failure candidate, not runtime or semantic completeness.

Witness fragment
witnessId: coupon-hidden-payment-cache
layer: static
kind: hidden_interaction
extensionClassification: non_split
measurementBoundary: measuredNonzero

Semantic witness

If rounding order changes the observable result, a selected semantic diagram can become non-fillable. The witness is tied to the selected diagram and measurement boundary.

Semantic non-fillability
apply coupon then round
  is not equal to
round then apply coupon

Report reading

A Feature Extension Report can summarize split status, introduced obstruction witnesses, repair suggestions, and non-conclusions. The non-conclusions are part of the report, not boilerplate.

  • splitStatus Whether the feature extension is split under the selected observation boundary.
  • introducedObstructionWitnesses Measured witness candidates that explain why the extension may fail a law.
  • repairSuggestions Advisory candidates such as introducing a port or moving access behind a contract.

Dynamics reading

A good coupon example can act as a seed attractor, while a bad shortcut can become a shortcut basin. This is an empirical / tooling reading of future change pressure, not a causal theorem.

Dynamics examples guide review and hypothesis formation. They do not prove global flatness preservation or future defect causality.