The User Experience

This section details the experience when using the speech-enabled Emacs W3 browser within the context of the audio desktop provided by the Emacspeak environment.

Audio Formatted Output

Well-structured WWW content is rendered using the technique of audio formatting introduced in AsTeR (Audio System For Technical Readings[1]). When using the default speech style sheet, Emacspeak renders structural components of the document, e.g., section headings at different levels, as well as hyperlinks in distinctive voices, thereby obviating the need to cue the listener to the document structure using added verbiage such as hot link.

Auditory Icons

Emacspeak uses auditory icons --short snippets of non-speech sounds-- throughout the audio desktop to provide succinct aural confirmation in response to user actions. In the context of the WWW browser, auditory icons are used to:

Fill Out Forms

As described earlier, Emacspeak and W3 provide a fluent speech environment for interacting with fill-out forms on the WWW. This section outlines the speech feedback a user hears when traversing standard dialogue elements such as checkboxes, radio button groups and pull down lists. The following list enumerates each user interface element along with the messages spoken by the system when traversing or manipulating these elements. For each interface element, we show a template for the utterance, and where necessary a concrete example where the template has been instantiated.
CheckBox
Simple Toggle
Traverse
Option option name or label is checked

Option search all pages is checked.

Toggle State
Unchecked search all pages
Radio Button
Mutually exclusive selection
Traverse
If button is pressed:

Group group caption if available currently set to button name or label.

Group kind of coffee set to Expresso

If the button is not pressed:

Press this to select Expresso as the kind of coffee

Press
Change state of group When a radio button is pressed, the user hears the sound of a button being pressed, and a spoken message indicating the updated state of the radio group.
Pull Down List
Select one or more elements
Traverse
List name or label currently set to value

List how many hits currently set to 100

Change selection
Selecting from a set with completion. System speaks current selection if any, and optionally prompts for the possible choices. The user can type ahead to interrupt speaking of the choices, and the system provides incremental completion and also speaks the list of matching choices based on what is typed.

The presence of the label enables Emacspeak to produce user friendly messages. In the absence of this semantic encoding, the system speaks the name of the field. This fall back strategy can work well when the name assigned to the field in the HTML encoding is meaningful; but this is often a function of the CGI script that processes the form. Typically, the name of a form field is the variable in the CGI script associated with that field; as a consequence, without the presence of the label, the speech user is left at the mercy of the variable naming convention used by the author of the server-side script that processes the form.

Accessing Information Efficiently

The tight integration of Emacspeak, W3, and the rest of the Emacs environment makes for a pleasant user experience; accessing WWW information that is being referred to in an email message or news article is only a keystroke away. This section enumerates a few simple user scenarios to give the reader a taste of the power available in this environment:
T. V. Raman
Email: raman@adobe.com
Last modified: Tue Feb 18 15:46:47 1997