Canonical scoring

What it does

• Tool normalization

  Runs axe-core, Lighthouse, Siteimprove, WAVE, and Pa11y against the same ScanEvent from the multi-domain scanner and reconciles their disjoint rule taxonomies into a canonical schema.

• Canonical score

  One number per domain (a11y, sustainability, AI Act), computed deterministically. The same input snapshot always produces the same score — no proprietary blackbox weighting.

• Cross-tool consensus

  Findings reported by multiple tools cluster as a single high-confidence violation; findings unique to one tool are flagged as lower-confidence and tagged with the source.

• Reproducibility

  A third party who runs the same scan on the same URL with the same rule-pack version computes the same number. No "ARIADA-magic" component; the algorithm is published.

• Signed measurement report

  Each per-domain score is emitted as a signed measurement report — cryptographic chain of custody from input URL to output number. Not a legal certification.

• Tool-sprawl collapse

  Replaces the typical 3–5 a11y tool stack and the 1–2 FTE integration team that babysits it. One number, one chain of custody, one signed report.

Layer mapping

Differentiator

Customers today wire together axe-core for CI, Lighthouse for performance, Siteimprove for the executive dashboard, WAVE for designer review, and Pa11y for headless probes — five rule taxonomies that disagree. Canonical scoring collapses that disagreement into one canonical score with cross-tool consensus. Reproducibility is the hard part: any third party running the same scan on the same URL with the same rule-pack version computes the same number. We never call the output a "certification" — that word belongs to accredited notified bodies.

Filed IP

ARIADA holds filed-IP positions covering the cross-tool canonical scoring and per-domain reproducibility methods at the core of this module. Provisional application only; conversion to non-provisional is scheduled within the 12-month window. PCT national-phase decision pending.

Application numbers, claim counts, and PCT deadlines are available for accredited-investor due diligence on the Legal & IP page.

Cross-references

• Multi-domain scan

  Source of the unified ScanEvent that canonical scoring consumes. Without it, the scorer would have to re-scan per tool. See multi-domain scan →

• AI authorship attribution

  Per-finding AI authorship is joined onto the canonical score so customers can see which fraction of their score regresses from AI-generated code. See AI authorship attribution →

• AI artifact audit

  Signed measurement reports are hashed into the tamper-evident evidence chain for AI Act Art. 50 packages.

• Regression detection

  Per-deploy canonical scores feed the regression detection diff engine to surface regressions cluster-by-cluster and per component.

• Trust page

  Methodology and the explicit boundary between "measurement report" and "legal certification" lives on the Trust page.

• Platform overview

  How canonical scoring snaps into the integrated pipeline alongside the other modules. See how the modules fit together →