Homer is described as a small self-voicing web browser for blind people. One of the goals of the whole Homer system is, "Kalliope," a long-term ambition to become a specialized Web portal for persons who are visually impaired or blind in Slovenia.
Homer is described as a, 'small, self-voicing web browser for blind people,' created by the Laboratory of Artificial Perception, systems and Cybernetics in the Faculty of Electrical Engineering at the University of Ljubljana in Slovenia. Homer was built for the GTK web browser, 'Dillo,' a free software project under the GNU general public license. The new, built-in screen reader is triggered by pointing the mouse and uses a text-to-speech module for output.
Homer has a dialog module which, together with a spoken-command input, was also introduced into the browser. It may be used for navigational purposes through web pages. The accompanying developed browser is mainly intended for use with the new web portal that is dedicated exclusively to blind users. The entire web portal, or sites that are linked from the portal, are arranged as common HTML/XML pages, complying with basic recommendations set by the WAI.
The last decade has found considerable advances made in the development of automatic text-to-speech and speech recognition systems. These systems offer more natural and user-friendly means of communication for persons who are either visually-impaired or blind. The goal of communication can be achieved more quickly and through access via computer networks, scanners and more that are available at fairly low costs.
At this time there are a number of research and development efforts under way to create Voice Browsers for the Web. Such browsers would allow any telephone to be used for accessing appropriately designed web-based services. The browsers would be usable by persons with visual impairments, or those who need web access while keeping their hands and eyes free for other purposes. The World Wide Web Consortium (W3C) has been the leader in these activities.
For a period of twenty years, spoken language technologies have been the main research area of activity at the University of Slovenia. Their prime interest is to develop a core technology for the Slovene spoken language that could be customized for different forms of applications. They have found the development of a small, self-voicing web browser for Slovene speaking persons with visual impairments or who are blind to be important to their research for a number of both technical and non-technical reasons; one of these reasons is the potential to assist persons with disabilities.
They decided not to follow the W3C guidelines and specifications while developing Voice Browsers. Their main interest is to develop a small, self-voicing web browser designed for blind users for accessing common web pages. The University of Slovenia believes that such a browser would never be used over a telephone line and that it needs to have a mouse-driven screen reader as well.
Homer was not built from scratch; new modules were introduced into the source code of publicly available browsers. The Homer system consists of five main modules, one of which is the original Dillo Web Browser. Input to the system is achieved via a keyboard with some specially-selected keys, or through using the speaker-independent spoken command recognition module. The dialog module manages the text-to-speech module. The dialog module manages dialogs with users, as well as performing access to web pages through the web browser module. Output from the system is written in a non-tagged, plain-text format using one of the standard character encodings, such as the Slovenian version of the seven-bit ASCIII coding or the WIN-1250 and the ISO-8859-2 codings. The screen reader module is optionally triggered by screen pointer motion events and uses the text-to-speech module as a means of output.
The Homer Browser was designed for use on a standard PC with a minimum of 64 MB of memory with a built-in, standard 16-bit sound card and a standard headset with a microphone. While it was initially developed for Linux, it was later ported to Microsoft Windows 9X/ME/NT/2000/XP as well. Homer performs best using multi-threading and additional advantages of a 32-bit operating system, and requires an additional ten megabytes of disk space for the program code and for the text-to-speech recognizer module inventory.
One of the goals of the whole Homer system is, 'Kalliope,' a long-term ambition to become a specialized Web portal for persons who are visually impaired or blind in Slovenia. Kalliope is planned to retrieve the EIS of the ZDSSS. All of the Web pages at Kalliope will comply with the basic recommendations set by the WAI, with a few additional XML tags, and will enable user-friendly navigation using the dialog module. The portal will serve as a site that links to other sites in Slovenia that are important to persons who are blind or visually impaired as well which are accessible to Homer.
Both Homer and Kalliope are in progress. The University of Slovenia expects the system to evolve towards a specialized web browser with a mouse-driven, text-to-speech screen reader and a voice-driven dialog manager that handles all of the web pages that are arranged at the Kalliope portal and sites linked through it.