Max Design

Published:

When we work with accessibility tools, we naturally build mental models to help us understand what’s happening under the hood.

Many of these models are reasonable, widely shared, and even reinforced by browser tooling and documentation. But some of them are also a little too neat.

Below are three accessibility assumptions that can quietly shape how we test, debug, and interpret results.


1. The DOM we see in DevTools is the real DOM

The assumption

The DOM shown in browser DevTools is the actual DOM. It represents a stable tree of elements that reflects what the browser is working with at any given moment.

The reality

For a deeper dive


2. The accessibility tree we see is an accurate picture of what assistive technologies receive

The assumption

The accessibility tree shown in browser tooling accurately represents what is exposed to assistive technologies, and inspecting it tells us what screen readers and other AT will perceive.

The reality

For a deeper dive


3. Screen reader output is a clean, ordered sequence of events

The assumption

Changes in the DOM or accessibility tree result in predictable, ordered announcements from screen readers, with each event mapping cleanly to a spoken output.

The reality

For a deeper dive


Why is it important to unpack these assumptions?

The DOM and accessibility tree are reliable guides most of the time, until you hit the edges where they stop being authoritative.

Understanding the nuances of the DOM, accessibility tree, and speech queue can help when testing and interpreting results.