WebInclusiveby Giuseppe Stumbo

Desktop screen readers

Try navigating a website, reaching a control and completing an action with a screen reader. Here is how to prepare and compare tests with NVDA, VoiceOver on macOS and JAWS.

Desktop checks with NVDA, VoiceOver and JAWS

Listen to how content and controls are read, then observe what happens when you use them. Connect names, structure and DOM order to the difficulties you encounter.

  • NVDA

    Desktop checks with NVDA

    With NVDA, you can follow the page structure, move between elements and use controls on Windows. Distinguish document reading from interaction with fields, then check dynamic messages as well.

    • Check tables, error messages, live regions, alternative text and icon-only buttons by listening to the announced name, role, state and value.
    • Compare at least the critical flows using a keyboard without a screen reader to determine whether the issue concerns focus, semantics or JavaScript behaviour.

  • VoiceOver

    Checking macOS with VoiceOver

    With VoiceOver on macOS, try finding your way around the page, choosing an element and completing an action. Hierarchy, groups, the rotor and Quick Nav help you observe how complex components work in Safari.

    • Test in Safari when the target requires it: use VO+U for the rotor and check headings, links, forms, landmarks and other available groups.
    • Check group navigation, entering and leaving components, focus order and consistency between VoiceOver focus and visual focus.
    • Use Quick Nav when needed to try faster navigation and identify patterns that work only with linear reading.
    • Check dialogs, menus, disclosures, tabs, tables, error messages and context changes without losing orientation.
    • Record the command, announcement, element and expectation: without these details, the bug remains difficult to reproduce.

  • JAWS

    Windows sessions with JAWS

    When a project requires JAWS, test the agreed journeys using the reader's modes. Check how headings, regions, fields and tables are read and how components are operated, recording the details needed to repeat the test.

    • Test the most critical journeys in Chrome or Edge by navigating headings, regions, form fields, buttons, links and tables.
    • Check forms mode, field names, instructions, errors, control states and custom component behaviour.
    • Check typical enterprise flows: login, filters, search, modals, data tables, exports and multi-step journeys.
    • Compare results with NVDA when you need to distinguish a site bug from a reader or browser difference.
    • Report the JAWS version, browser, command, announcement and impact on the task.

Modes of use to distinguish on desktop readers

Repeat the journey in the modes provided by the reader: structural reading, focus within controls, groups and shortcuts can reveal different behaviours. Record which mode you are using.

  • NVDA

    Browse mode and focus mode

    With NVDA, always distinguish reading in browse mode from focus mode on interactive controls: many widgets appear correct in only one of the two.

    • Use H for headings, D for landmarks, K for links, B for buttons, F for form fields and Shift+key to return to the previous element.
    • Open the elements list with NVDA+F7 to check headings, links and landmarks without scrolling through the whole page.
    • Switch between browse mode and focus mode with NVDA+Spazio when entering fields, menus, tabs, dialogs, comboboxes and autocomplete controls.
    • Check whether the reader clearly announces entering, leaving and the states of custom components.

  • JAWS

    Virtual reading, forms and keyboard shortcuts

    Test JAWS both for general page reading and for enterprise interactions, where shortcuts, forms mode and legacy compatibility still matter greatly.

    • With the virtual cursor, use H for headings, R for regions, T for tables and F for form controls; U and V reach unvisited and visited links respectively.
    • Check transitions into forms and custom controls, particularly when focus changes dynamically.
    • Compare results with NVDA and inspect the semantics exposed by the browser: a difference between readers helps guide investigation but does not prove the cause on its own.

  • macOS

    VoiceOver with rotor, groups and Quick Nav

    On macOS, reading content is not enough: find out whether groups, the rotor and Quick Nav actually help people move through the layout and interactive components.

    • Open the rotor with VO+U, move between headings, links, forms and other structural elements, then navigate the list.
    • Check group navigation and moving into and out of complex components or composite areas.
    • When needed, also check Quick Nav to see whether the flow works without sequential exploration.