When text is not written through the operating system or through a documented API, assistive technology, such as screen readers, cannot render the content to the user. The first example below shows a form with several text input fields. In this example, the fields are cross-platform widgets that don't expose the text input location. In this case, a screen reader would likely announce the values "Jane" and "Doe" as a whole.
While the above scenario provides some accessibility information, the form fields are not fully accessible. To demonstrate, below and to the left is the same form; now the focus caret has shifted to the third character, "n". In the case of a cross-platform control, the screen reader does not announce the letter "n", and the user is unaware of what character he or she is on. Without knowing the character, the user would be unable to edit the input field to change the value.
Similarly, in the example below and to the right, text has been highlighted. When text is not written to the operating system or through an API, the screen reader user would not know what text is selected. If, for example, the user selected text to copy, he or she wouldn't know if just one character or the entire word were selected.
Select Next to learn about testing for this requirement.