Abstract: Web pages and other text documents displayed on a computer are reformatted to allow a user who has difficulty reading to navigate between and among such documents and to have such documents, or portions of them, read aloud by the computer using a text-to-speech engine in their original or translated form while preserving the original layout of the document. A “point-and-read” paradigm allows a user to cause the text to be read solely by moving a pointing device over graphical icons or text without requiring the user to click on anything in the document. Hyperlink navigation and other program functions are accomplished in a similar manner.

Abstract: User interaction of a visually displayed document is provided via a graphical user interface (GUI). The document includes, and is parsed into, a plurality of text-based grammatical units. An input device modality is selected from a plurality of input device modalities which determines the type of input device in which a user interacts with to make a selection. One or more grammatical units of the document are then selected using the selected type of input device. Each grammatical unit that is selected is read aloud to the user by loading the grammatical unit into a text-to-speech engine. The text of the grammatical unit is thereby automatically spoken. Furthermore, a switching modality is selected from a plurality of switching modalities. The switching modality determines the manner in which one or more switches are used to make a selection.