Improving text accessibility and understanding of domain-specific information

Abstract

This thesis explores ways to make text more accessible for end users, taking into consideration the multiple layers of text. Most attention is given to the specialized language layer, where lay users encounter issues while accessing domain-specific information. Techniques and tools for accurately identifying and explaining specialized terminology, even if present in derivate form, are presented. The impact of this process is evaluated on medical language and results reflect an improvement of message understanding for lay persons. Additionally, means to adapt the text at presentation level are presented, giving special attention to users with dyslexia or low vision. The resulting tools were designed with careful attention to availability and usability, making them run anywhere with minimal effort. Integration of the developed services within a model of universal text accessibility is also presented.