| Abstract | Identifying accessibility issues that can threaten universal website usability is criticalfor web service and content providers who wish to accommodate the diverse web audience.
Detecting page-to-environment incompliance and modifying pages to promote universal
accessibility is one important step in improving the process of exploration and navigation
in the web user experience. To address this issue, we have designed a system that evaluates
the accessibility of a web page in a given browsing environment based on knowledge of
the HyperText Markup Language (HTML) tags that comprise the page and knowledge
of the tag support provided in respective browsing environments. Given this approach,
one of the most important aspects of the system is the comprehensive nature of tag support
knowledge. The more support rules known, the more environment-specific bugs the system
can accurately identify. In order to optimize knowledge of tag support criteria, we have
also incorporated a learning mechanism that can inductively determine HTML tags that
are unsupported in a given environment by observing both positive and negative examples
of web page appearance and behavior.
|
