Required outcomes
The registry keeps all 55 A and AA success criteria in one place, each with an original operational summary and links to the normative text and its Understanding document.
LAKA–WCAG Accessibility Grammar · version 1.0.0
Accessibility, written as grammar.
WCAG 2.2 defines the required outcomes. The LAKA model defines diagnosis, repair depth, measurement and feedback. This grammar joins the two: every one of the 55 Level A and AA success criteria gets an operational summary, and every assertion is written as explicit IF / THEN / ELSE logic with declared evidence, owners and exceptions.
A rule is incomplete until its object, scope, states, required outcome, source, evidence method and owner are all named. A screenshot alone never establishes interaction behaviour.
What is in the package
55
WCAG 2.2 A/AA criteria
50
LAKA base-grid cells
14
Measurement lenses
700
Addressable analysis cells
Counts derive from the corpus at load time: 31 Level A and 24 Level AA entries (the removed 4.1.1 is excluded), 6 schema-validated example rules, 71 primary sources and 11 documents — all served at /v1.
The split of authority
WCAG says what. LAKA says how deep.
Diagnosis & repair depth
The LAKA foundation crosses five change levels with ten internal variables — 50 base cells that decide how deep a repair goes — and reads every change through 14 measurement lenses.
Explicit logic
The grammar writes each assertion as IF / THEN REQUIRE / ELSE with AND, OR and NOT combinators, three separated authority layers, and evidence semantics that refuse to guess.
Fourteen lenses
How a change is measured
Every accessibility change — a repair, a regression, a redesign — is read through the same fourteen questions.
Magnitude
How much does this aspect change?
Count changed properties and affected task steps; report barrier reduction separately.
Rate
How quickly does it change?
New confirmed barriers per release or week, using a fixed scope.
Direction
Does the change improve or worsen independent access?
Trend in task barriers and successful completion; name the desired direction.
Scope
How broadly does it apply?
Unique affected routes, roles, components and journeys.
Depth
At what system layer does it occur?
Instance, component, shared primitive, architecture or product model.
Duration
For how long does the condition or barrier persist?
Time present per interaction and age of unresolved finding; label which is measured.
Frequency
How often does it occur?
Affected encounters divided by observed eligible encounters.
Acceleration
Is its rate increasing or decreasing?
Change in barrier-arrival rate across comparable time windows.
Variability
How different is behavior across contexts?
Differences across themes, viewports, locales and input/assistive-technology combinations.
Detectability
How reliably can it be found?
Detection method and, when measurable, pre-release discoveries divided by all subsequently known findings for that release.
Reversibility
Can the change or action be safely undone?
Verify user recovery and deployment rollback separately; record successful trials.
Propagation
Where can a local change spread?
Trace consumers and descendants of the changed primitive or content field.
Amplification
How much additional impact can one source create?
Downstream task failures attributable to one source defect, without double counting.
Accumulation
What builds up over time?
Unresolved confirmed backlog and its change over comparable releases.
Read this before you use it
A specification, not a scanner.
A specification and reusable content set — not an installed accessibility scanner, legal opinion, certification or completed website audit. All sample observations are explicitly untested. WCAG 3 remains draft guidance rather than the target for this package.