Accessibility

Design criteria

We aim for semantic structure, keyboard navigation, visible focus, sufficient contrast, text alternatives and understandable content. Visualizations should retain an accessible HTML alternative.

Report a barrier

If you encounter a barrier, tell us the URL, the task you attempted, the browser or assistive technology used and the expected result. We will review the issue without requesting unnecessary data.

Testing and continuous improvement

WCAG 2.2 informs our work on perceivable, operable, understandable and robust experiences. We combine automated checks with keyboard and visual review, paying particular attention to forms, focus order, zoom, reduced motion and interactive visualizations. A reference to this standard is a direction for improvement, not an unqualified conformance claim.

Interactive content and media alternatives

Astronomical maps, scale comparisons and animated explorers must not be the only route to their information. We provide headings, explanatory text, lists or tables that preserve the essential facts outside the visualization. Controls need accessible names, keyboard operation and visible focus; motion should respect reduced-motion preferences and must not trap normal page scrolling. Features that cannot work reliably on small touch screens may be replaced by a static, accessible alternative.

Meaningful images receive alternative text appropriate to context. Decorative elements are hidden from assistive technology. Video and audio should provide captions, transcripts or an equivalent summary when they carry editorial information, while color alone is not used to communicate status or meaning.

Review, limitations and prioritization

Automated accessibility checks are useful for regressions, but they cannot judge reading order, clarity, alternative-text quality or whether an interaction is practical. We therefore combine them with keyboard navigation, zoom and contrast review, responsive testing and inspection of names, roles and states. Reusable components are preferred because one verified improvement can benefit many pages.

Reported barriers are prioritized by impact: inability to reach content or complete a form comes before a minor visual inconsistency. We record the page, environment, expected result and reproduction steps, then add regression coverage where practical. Known limitations should be described honestly; this page is a working policy, not proof that every page is free of barriers.

Accessible editorial language and data

Accessibility also applies to the writing itself. We expand uncommon abbreviations, explain specialist terms at the level needed by the story, use descriptive headings and avoid making readers infer relationships that can be stated directly. Tables need meaningful headers; units, symbols and dates must be unambiguous; links should describe their destination rather than rely on repeated phrases such as “click here.” Translations preserve meaning and structure instead of mechanically mirroring word order.

Dense scientific material can offer a concise explanation before technical detail without removing precision. Mathematical notation, coordinates and very large scales should be accompanied by a textual interpretation where practical. Error messages identify the field and corrective action, and status changes are announced in more than color alone. These editorial choices benefit screen-reader users, people with cognitive disabilities and readers unfamiliar with the subject.