Challenges of voice-enabling Web content

Abstract

Mainstream Internet usage is growing ripe into a pervasive set of services people use as common tools in everyday life, like cellular phones and a sort of different wearable gizmos. Voice-based access enhances Web accessibility by enabling the physically and visually impaired community. Unfortunately, ubiquitous browsing lacks a consensual delivering format which overcomes the physical constraints, memory and processing limitations of such wearable gizmos. In addition, delivering to such devices often occurs under time-pressure and content should be customized to user and context. For all these reasons, hypertext looses much of its benefits. In this paper we discuss the main differences between visual and voice-based access to Web content and the challenges of providing customized vocal interfaces to intrinsically visual oriented Web applications. This paper also provides some guidelines towards a solution to the problem. The key ideas concern the use of natural language generation, voicexml and Web services.