Design and validation of a user-centered Web browser
Abstract
Human cognitive limitations in memory, attention, and decision-making in using Web browsers were analyzed, and factors involved in the design of Web browsers were investigated. A three-dimensional model is presented to predict the relationship among search strategies, browser types, and task types. A conventional and an enhanced browser were built by rewriting the interface of Netscape Navigator. The enhanced browser included the following features: shopping list, pull back, preview information, and reference window. JavaScript and Common Gateway Interface (CGI) programming were used in the software development. An experiment was conducted to test four hypotheses concerning the validity of the model and the usability of the enhanced browser. General-browsing and item-searching tasks were used with 64 subjects. The dependent variables were the mean path length, the identified item number, the number of steps per identified item, and the satisfaction rating. The independent variables were task types and browser types. A cross-nested design was used, within-subject for task types, and between-subject for browser types. Testing of the four hypotheses indicated the following: (1) In the general-browsing task continuous strategies were used, while in the item-searching task, discrete strategies were utilized; (2) In the item-searching task, the subjects' strategies used in the enhanced browser were more discrete than those in the conventional browser; (3) Once the subjects became skilled at using the enhanced browser their performance was superior to using the conventional browser; and (4) There were no difference in satisfaction between using the enhanced and conventional browsers. However, in the enhanced browser, the subjects were more satisfied in using the reference window and the shopping list, and less satisfied in using the pull back and the preview information than using the conventional browser. The restrictions inherent in the experimental design did not allow to fully exhibit the true potential benefits that may accrue with the use of the new features.
Degree
Ph.D.
Advisors
Salvendy, Purdue University.
Subject Area
Industrial engineering|Computer science|Occupational psychology
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.