Accessibility
What we have built, what has actually been tested, and what has not. The last of those is the part worth reading.
- Pages are structured the way they read — headings in order, real lists, tables with proper headers, and every form control labelled.
- A “skip to content” link is the first thing on every page, so a keyboard or screen reader can jump straight past the navigation.
- Visible focus indicators everywhere. Focus outlines are never removed, only restyled.
- The section of the site you are in is marked three ways at once — announced to a screen reader, and shown by weight and an underline — so it never depends on colour alone.
- Colour is used for meaning in only two places, and both always carry a text label as well.
- Full keyboard operation. There are no drag interactions, no hover-only controls and no custom widgets that trap focus.
- Calculator results and search results are announced to a screen reader when they change, rather than appearing silently.
- Wide tables scroll inside their own container; the page never scrolls sideways.
- If you have asked your device to reduce motion, we honour it — transitions and smooth scrolling are switched off.
- Dark mode follows your system setting, with contrast checked in both.
- Decorative elements, including the rating bars, are hidden from assistive technology because their meaning is already in the adjacent text.
Separately from the site, every activity we write up carries notes on how accessible the activity is — barriers, adaptations, and the adapted forms that exist. We treat that as editorial content rather than a compliance exercise, and we write it as prose rather than a score, because “how accessible is it” is never the real question. The real question is “accessible to whom, and with what”.
Every time the site is updated, an automated accessibility checker runs over the finished pages as they are actually published — not over pieces of them in isolation — covering one representative page of every kind. It checks page structure and heading order, whether every control and link has a name that makes sense on its own, whether labels and languages are set, and whether lists and tables are properly formed. A single failure stops the update going out.
No screen reader test has been performed. Not with VoiceOver, NVDA, JAWS or TalkBack. The pages are constructed to be read aloud well and nobody has yet sat and listened to them, and those are different things.
Colour contrast is reasoned, not measured in place. The automated check inspects the pages without ever drawing them, so it cannot see the colours as they finally appear on screen. The palette was chosen against the contrast thresholds, but that is a decision we made rather than a measurement of the published pages.
Tap-target sizes have not been measured. The same limitation: it needs the pages drawn at a real size on a real screen.
No conformance claim is made. We have not audited this site against WCAG at any level, and saying “WCAG 2.2 AA compliant” without having done that audit would be a claim we cannot support.
If something here does not work for you, that is a fault on our side and we want to know about it — particularly anything involving a screen reader, since that is the largest gap we know about.
See also about HobbyHQ and privacy.