Automated Usability Evaluations using Client Side Usability Framework (CSUF)

Abstract

Usability of a website is essential in order for a company to achieve its business goals. Evaluating a website for usability and accessibility is costly and time consuming primarily because it involves usability experts and users. Automated usability evaluation methods try to alleviate this problem by automating certain aspects of the usability evaluation. One kind of these automated methods evaluates the usability by checking the HTML composition of a webpage against a pre-defined set of research-based guidelines. With only a fraction of the webpages developed being static, these methods cannot be used on modern webpages and applications. In this thesis, we have developed a client side usability framework (CSUF) which addresses this problem. By encoding the guidelines as JSON and fetching the HTML elements dynamically for evaluation, we have observed that websites can be evaluated for usability irrespective of the web programming language used and irrespective of the design pattern used. Furthermore, the number of guidelines to be encoded were reduced using CSS selectors efficiently and the developer retains the ability to perform the evaluation at any point of page rendering or modification.