Seminar 2: Reflection

Seminar 2

In Interaction Design: Beyond Human-Computer Interaction, chapter 13, we discuss “An Evaluating Framework” where we get introduced to alot of concepts and thoughts on how to do proper evaluation. The first one that pops up early on is the DECIDE framework. It is a concept that resembles a checklist to help plan an evaluation study. It also helps reminding about the issues that might stem from before or during a study. The chapter in general brings up some important aspects such as participants, facilities and equipment, expertise and validity.

In chapter 15, heuristics start to come up. Heuristics is studying or predicting that a system follows a set framework and conform to tested principles. Here we finally get to follow the true meaning of HCI where the writer mentions key points of parts of the system that has to work. This is called Heuristic evaluation and was developed by Nielsen. It’s in short 10 different points of usability, whereas an example of one of these points could be that the system always should keep users informed about what’s going on through appropriate feedback wihin reasonable time. (Belongs to the first point Visibility of system status). Aesthetics, flexibility and error prevention are also mentioned along with many others. These serve in my opinion as a great foundation for developing a system and also are of utter importance to follow.

Moving on, hedonic heuristics originating from Finland is a new kind of heuristic that directly addresses how users feel about their interactions. To elaborate, you could ask the question “Does this system satisfy users curiosity by offering interesting content frequently?”. This can be reflected in many websites today that are pretty obvious when you look closer. An example of this would be Apples website where they offer pleasing aesthetics and interesting content. They present to you a minimal but yet necessary interface and filling every page with interesting content.

Heuristic evaluation has three stages, the briefing session, the evaluation period and the debriefing session. The first one is what it sounds like, everyone gets the same briefing. The second phase is where each expert spends 1-2 hours inspecting the product and using the heuristics as guidance. The last phase everyone comes together and discuss their findings, prioritizing problems they found and suggest solutions.

Question for the seminar: How important is it to follow these heuristics when developing a product? How do you choose which model you use without hiring an expert?

Extra (Explaining some of the walkthroughs and models, summarising for future reference for the project. > 400 words):

In contrast, cognitive walkthroughs focus more on identifying specific user problems at a high level of detail.  Will users know what to do? You answer this question with either yes or no, focusing exactly on why not. Both the cognitive and heuristic are very usable in different environments and can both serve as guidelines on different places of the products timeline.

Pluralistic walkthroughs focus on experts and developers assuming a role of a user to simulate in a way each step through a scenario of a product. They then discuss these tasks at a detailed level and unfortunately the biggest hindrance of this type of walkthrough is either a time constraint or that all experts are together in one place, making progress seemingly slower.

Following up the inspection methods and all the analytics, the predictive models basically evaluate a system without users being present. The model uses formulas to show measures of user performance and can be seen as the more mathematically sound technique. In other words, predicting the amount of time it takes for a given task and the most efficient strategies to use.

One of these techniques is called GOMS model along with its offspring, the keystroke level model (KLM). The first one sets up a goal, a method (learned procedure to accomplish the goal), and being described by the operators. The writer of the book gives an example of the goal to delete a word in a sentence, where the method is each step in doing so. The operators in the method are the general description of the methods, in other words the execution.

The latter model I mentioned, the KLM method, to juxtapose with the GOMS is that it provides exact numerical predictions of user performance. Pressing a single key on a keyboard is for example predicted to take 0.35 seconds for a user. Pointing the mouse to a target on the display approximates to 1.10 seconds. Using this method gives a concrete and solid basis for calculating the amount of time a given task can take for a regular user. One of the cons with this model is that it doesn’t take a users muscle memory, self-aware users or even design of a system into consideration with these numbers. Writing today on a keyboard takes a very skilled writer alot less time then 0.35 seconds per keystroke. This would mean that this text would take just over 30 minutes to write in just pressing keys.