Comber, T 2010, 'An analysis of entropy, redundancy and complexity in graphical user interfaces', MBA Southern Cross University, Lismore, NSW.
Copyright T Comber 2010
This thesis introduces a new sequential data analysis technique for graphical user interfaces (GUIs) based upon the communication theory of Shannon. This study addresses two research questions:
Research Question 1: Can an entropy analysis of a sequence of operations be used to estimate the quality of the user‟s work?
Research Question 2: Can an entropy analysis of a sequence of operations be used to detect usability problems?
Shannon‘s entropy equation is applied to a word processor ―language‖ for 24 users and a language entropy value is obtained for each. This value is correlated with metrics associated with experience, skill, knowledge and percentage of task completed. The results show a sufficiently high correlation to justify the use of language entropy as an independent user competencey and/or application usability metric. It is suggested that new GUI applications be built (or existing ones patched) to provide an automated entropy ―score‖ for user sessions. The advantage is that a user‘s language entropy can be automatically determined during a normal work session, negating the need for costly, labour intensive and time-consuming methods such as interviews, questionnaires and other usability protocols. An argument is presented to (a) incorporate the entropy score into user-centred design and evaluation sessions as a new usability metric, and (b) use the score as a measure of the effectiveness of user training and the degree of user competence when undertaking GUI software tasks.