Web Interaction Environments: Characterising Web Accessibility at the Large

• Autor(a) principal: Lopes, Rui Miguel do Nascimento Dias
• Data de Publicação: 2012
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10451/14309
• topic: Web Accessibility; Web Characterisation; Web Science

Accessibility quality on the Web is essential for providing a good Web experience to people with disabilities. The existence of virtual ramps aid these users grasping and interacting withWeb content, just like the experience of those who are unimpaired. However, more often than not, Web pages impose accessibility barriers, usually centred on the unavailability of tailored content to specific perceptual abilities (e.g.,textual description of images, enabling grasping information with assistive technologies), as well as on proper HTML structural elements that adequate the semantics of a Web page. When evaluating the accessibility quality of Web pages, the resulting analysis is often focused on a small sample set (e.g., a single Web page or a selection of pages from a Web site). While this kind of analysis gets the gist of accessibility quality, it misses the big picture on the overall accessibility quality of the Web. This thesis addresses the challenge of observing accessibility phenomena on the Web, through the experimental evaluation of large collections of Web pages. This resulted on new findings about the accessibility quality of the Web, such as its correlation with HTML element count, and the erroneous perception of accessibility quality by developers. Small-scale experiments have been verified also at large scale, such as the correlation between the usage of HTML templates and accessibility quality. Based on the challenges raised by the experimental evaluation, this thesis proposes a novel approach for large scale Web accessibility evaluation based on Linked Data, as well as the establishment of metrics to assess the truthfulness and coverage of automated evaluation methods.