Skip to main content
WCAGrules
Quick navigation

ACT Rules

How the rules are actually tested

A technique shows one way to satisfy a criterion. A failure shows one way to break it. An ACT rule is the third piece, because it defines exactly how you test for one, and it is the reference that automated checkers implement. We carry 87 of them here, each mapped to the criteria it tests and to our own fixes for those criteria.

The part worth reading twice is what a pass is worth. On most of these rules W3C's own answer is that passing means no failure of that kind was found and the criterion still needs further testing. Four rules in the whole published set clear a criterion outright, and they cover character key shortcuts, page titles, the language of the page and the language of its parts. So a green result here is evidence that one specific thing is not broken, and it is not evidence that the rule is met.

What a tool can and cannot reach

Of the 87 rules here, 56 can be decided by a tool on its own. The rest need a person somewhere, either because the check calls for judgment or because nothing has implemented it yet. Between them they touch 37 of the 86 WCAG 2.2 criteria. W3C's own index runs slightly ahead of ours and reaches 38, which still leaves 48 criteria in the standard with no test rule of any kind.

That is the honest shape of automated coverage, and it is as close to an answer as the standards body offers, because W3C has never published a figure for how much of WCAG a machine can check. Anyone quoting one is quoting somebody's classification rather than the standard. Ours is on what automated testing actually covers, and it is labeled as ours.

Where the rules cluster, and where they run out

Coverage is not spread evenly across the standard, and the shape of the gap says more than the total does. The rules pile up on the guidelines about markup a machine can read, and they thin out on the guidelines that ask whether a person understood something. So the count below is worth reading as a map of what automation reaches rather than as a score.

The 13 WCAG 2.2 guidelines, with the number of live ACT rules filed against each, the number of Level A and AA criteria it holds, and how many of those criteria a rule reaches.
GuidelineLive rulesCriteria at A and AAOf those, reached by a rule
1.1 Text Alternatives811
1.2 Time-based Media654
1.3 Adaptable1454
1.4 Distinguishable1196
2.1 Keyboard Accessible633
2.2 Enough Time322
2.3 Seizures and Physical Reactions010
2.4 Navigable1585
2.5 Input Modalities462
3.1 Readable522
3.2 Predictable250
3.3 Input Assistance161
4.1 Compatible1521

Two rows are worth stopping on, because both of them cost money in a way the totals hide.

Seizures and Physical Reactions has no rule at all. Not a thin one. None, across every criterion in it. A scanner therefore returns a clean result on the one guideline whose failure mode is a medical event rather than an inconvenience, and it returns that result honestly, because nobody has written a rule for it to run. The number that decides 2.3.1 Three Flashes or Below Threshold has to be measured by a person with a tool built for it.

Input Assistance holds nine criteria, six of them at A and AA, and one rule. That guideline is where forms live, and forms are where most sites take money. Its single rule is still at proposal stage, and 3.3.2 Labels or Instructions has nothing at all, which is why a field labelled only by placeholder text passes every checker and fails an audit. So the part of your site that earns is the part automation reaches least.

And where the rules are thickest, they are thick on one criterion. 4.1.2 Name, Role, Value is filed with fifteen live rules, more than anything else in the standard, and ten of those name it as a conformance requirement. Seven of the ten test only that an accessible name is not empty, so a button whose accessible name is the literal word button passes all seven. The other three are the ones doing real work, because they test focusable content hidden from the accessibility tree, focusable content sitting inside presentational children, and two iframes sharing one name for different purposes.

Rules by criterion

1.1.1 Non-text Content Level A

1.2.1 Audio-only and Video-only (Prerecorded) Level A

1.2.2 Captions (Prerecorded) Level A

1.2.3 Audio Description or Media Alternative (Prerecorded) Level A

1.2.5 Audio Description (Prerecorded) Level AA

1.2.8 Media Alternative (Prerecorded) Level AAA

1.3.1 Info and Relationships Level A

1.3.3 Sensory Characteristics Level A

1.3.4 Orientation Level AA

1.3.5 Identify Input Purpose Level AA

1.4.2 Audio Control Level A

1.4.3 Contrast (Minimum) Level AA

1.4.4 Resize Text Level AA

1.4.5 Images of Text Level AA

1.4.6 Contrast (Enhanced) Level AAA

1.4.10 Reflow Level AA

1.4.12 Text Spacing Level AA

2.1.1 Keyboard Level A

2.1.2 No Keyboard Trap Level A

2.1.3 Keyboard (No Exception) Level AAA

2.1.4 Character Key Shortcuts Level A

2.2.1 Timing Adjustable Level A

2.2.2 Pause, Stop, Hide Level A

2.2.4 Interruptions Level AAA

2.4.1 Bypass Blocks Level A

2.4.2 Page Titled Level A

2.4.4 Link Purpose (In Context) Level A

2.4.6 Headings and Labels Level AA

2.4.7 Focus Visible Level AA

2.4.9 Link Purpose (Link Only) Level AAA

2.5.3 Label in Name Level A

2.5.4 Motion Actuation Level A

3.1.1 Language of Page Level A

3.1.2 Language of Parts Level AA

3.2.5 Change on Request Level AAA

3.3.1 Error Identification Level A

4.1.2 Name, Role, Value Level A

Rules about using ARIA correctly

These check that ARIA is used the way the specification intends, rather than checking a WCAG criterion. W3C is explicit that some of them are not required for conformance at any level, so treat a failure here as a sign that something is probably wrong nearby, often around 4.1.2 Name, Role, Value, rather than as a finding that stands on its own.

Rules for audio and video

Each of these checks one piece of a media requirement rather than a whole criterion, and W3C feeds them into larger composite rules. So one of them passing tells you nothing about whether your captions or your audio description are adequate. Which criterion applies depends on what the media contains, whether that is audio only, video only, or both, which is what Guideline 1.2, Time-based Media sets out.

Withdrawn rules

These were published and have since been withdrawn. They are here because you may still meet one in an older audit report or a tool's output, and the useful thing to know is that it no longer stands.

Where to go next

Go somewhere useful

Find tools, resources and your workspace.

29 destinations

↑ ↓ navigate · Esc closeSearch the full site