Reading Seminar 2- Dawood Rehman

The aim of chapter 13 was to introduce the main key concepts and terms used in evaluation. The chapter emphasized the importance of evaluation and generalized different types of evaluations into three broad categories, controlled settings involving users, natural settings involving users and any settings not involving users. There are pros and cons of each category. For an example, lab-based studies are good at showing usability issues but they are poor at capturing the context of use while modelling and predicting approaches are cheap and quick to perform but can miss crucial information about unpredictable usability problems. I think that the best evaluation framework would be obtained from a combination of different evaluation methods that are derived a combination of the three broad categories presented. Our group should therefore assemble an evaluation framework that is feasible and consists of a combination of various methods that are derived from the three main categories for a richer understanding of problems that can arise.

Chapter 15 introduces various inspection methods, heuristic evaluation, walkthroughs, analytics and predictive models. These methods are unique in the sense that they do not require users to be present during the evaluation. I think that heuristic evaluation and walkthroughs will play a great role in the development of our prototype. We can use heuristic evaluation to evaluate whether user-interface elements conform to tried and tested principles. Using heuristic evaluation as an iterative process is really the key to getting a good high-level design. If we think that a particular part is very important and really needs to be evaluated properly we can use walkthroughs for that particular area since they are suitable for evaluating spall parts of a product. I think that analytics and predictive models will be somewhat hard to use because we don’t have an actual product which we can collect data about to improve the lagging areas and we don’t really have the time to create reliable and realistic predictive models unless we use already existing models such as Fitts’ Law.

Question: How can we assemble a feasible evaluation framework that will highlight and give us a richer understanding of the problems and flaws in our prototype?