In chapter 13 we are introduced to the concept of evaluating your prototype. They go through why, what, where and when you should do your evaluation. The "why" is pretty much self-explanatory. We need to evaluate our product prototype since we need to know if it will make it out on the market. What to evaluate depends on the main goal with the final product, but a product could have more than one main goal. For example a cell phone needs both to have a functioning software that is easy to use but also a good design so that people want to be seen with it. Where to evaluate depends on what is being evaluated. There is something called "in the wild studies" which means that the product is being evaluated in its natural environment. When to evaluate mostly depends on the product. If it is a new product on the market then the evaluation should take place after the requirements has been established and the first sketches has been done so you can see early in the process if the product is worth continuing working on.
There are three main types on evaluation. These are "Controlled settings involving users", "Natural settings involving users" and "Any settings not involving users". The names speaks for themselves what the different types are. For example a controlled setting could be in a lab and a natural setting could be in the jungle. The third types is usually when the creators try to predict what usability problems that needs improvement.
This principle can be applied to our project when we our done with our model and have our first prototype working so that we have something of value to evaluate.
The 15th chapter is an extension of chapter 13 and is also about evaluation. More specifically about evaluation methods not involving direct contact with users but more in a way of using heuristics to predict the users' performance. This is especially good when you don't have the ability to test your product in reality.
Heuristic evaluation is evaluation made by experts where they use different principles to check if the product fulfills these.
Question: Can a whole evaluation be done by not involving users?