...
Since our project was focused on giving users a forum to write, we wanted to design an interface that did not interfere with their work. Since the functionality required for each page was simple, we aimed to create a very simple site. This proved remarkably challenging, both practically and aesthetically. Streamlining the user actions required several levels of user testing and heuristic evaluation to design – at every stage, we found that our ‘simple’ arrangement of the buttons and links was confusing to some of the the testers, and would be better placed elsewhere. However, with a simple design as a goal, elsewhere was sometimes hard to find. Similarly, while the layout of our page appears very simple and understated, choosing colour and a layout that neither looked bare nor created unnecessary cognative cognitive dissonance took a very long time. In general, however, we did find that simple layouts were helpful for the users, as they could quickly see all the options and possible decisions.
If we were to create a new project, we would want more user input, and more frequent, finely grained input. We feel that the first problem discussed (that different users have different expectations), would be alleviated if the volume of user tests were increased, since we would then be able to determine which expectation was held by the majority of our prospective users. We also felt that recreating the entire design for every user test meant that user input was less frequent and less effective than it might have been. For example, our paper prototype had the most problems with the story page, and it might have been useful to have a quick mock-up of the new design for just that page (either in paper or as a computer prototype), without any real features, just to check that our new design fixed the problems of the old and did not create any new difficulties. Similarly, our task were very story-centric, so the the user testing illuminated problems with the story page, but not as many problems with the main page, since the test users did not interact with it extensively. As such, the heuristic evaluation provided a lot of useful feedback about the main page that we had not considered before. If there is some part of the design that does not play a large role in user testing, it would be nice to be able to get a heuristic on just that page early on in the project, so that any problems with it will not interfere with the testing of the other pages, and can be fixed quickly. Sending only that part of the site for testing would decrease the strain on the evaluators and the creators, hopefully allowing them to create the best possible interface.