WTG
People with visual impairments have a hard time using the internet, especially web pages. Persons with visual impairments usually use screen readers which reads out the web pages to them. The problem with web pages however is that there are too many non-content (headers, design, menus) that the screen readers will read. Therefore it takes a long time for the users to read a certain web page.
The research performed at the Health Information Management Department, University of Pittsburgh aims to rectify that problem. My main research assistant duties was to create a program that can transform web pages to something that can easily be read by screen readers. To do this, we first convert the website so that it becomes accessible according to the WAI standards of W3C. Furthermore we try to make the websites more accessible by detecting the various segments and rearranging them according to what is the most important task.
Architecture of the WTG program.
Contribution
My main contribution to this research was first of all to create the basic architecture for the system. The system was created using an object oriented architecture (which allowed for various object plugins) using JSP and java servlets.
Secondly, I introduced the concept of segment pattern matching, where the program tries to automatically detect segments and segment type based on the html characteristics of the text. Once these segments were correctly identified, we can move the various segments around accordingly
Technical Details
- Built using: JSP, servlets
- Time: August 2003-August 2005