Welcome back. The topic for this video is Conceptual Models. This video draws on work first stated by Donald Norman in his book, The Design of Everyday Things. Now the the idea of conceptual models is that we as people are explanatory creatures. We come up with models. We have a drive to figure out how things work. We need to come up with these models because we also need to use things in new ways. And so rule 1 of interface design, is you want to create an interface that behaves exactly as users think it will. Another way of putting this is your job, is to make it easy for users of your system to create the right model of the system. Now let's look at this in a little bit more detail. When you're a designer and you create a system, you have a model of how the system works. Or put it another way, you know how the system works. You know the insides of the system, because you built it. Now hopefully, that model is explicit. That's good design practice, good software engineering practice. And so you have your model of the system as a designer. But of course the user has to figure out how it works. The user has to build a model. And that is called the user's model of the the system. Now, we hope that the user's model will be identical to the designer's model of the system. That is, they will have an accurate understanding of how the system works. Well, how can the user actually create the model? They don't have access to your knowledge, they don't have access to the internal workings of this system. So the only way they can build a model of the system is through what's called the system image. All of the things about the system that are visible to the user. The controls, the displays, and so on. That is the system image. And so that is what the user uses to build a model and as a designer, your job is to create a system image that makes it easy for the user to acquire a user model that is equivalent to the design model. So let's consider an example, and as often is the case, an example that shows when this is not done well. Illustrates the importance of this concept, perhaps a little better than even a good example would. So, I'm going to give you an example of how to control temperature in a refrigerator. There's two things that you want to control about, when it comes to the temperature of a refrigerator. The temperature of the fresh food compartment, where you might put milk and orange juice and things like that. And then the temperature of the freezer compartment, where you put frozen food. And this example is adapted from, The Design of Everyday Things, by Don Norman. So here are instructions for how to control the temperature of a refrigerator, and notice that there's fairly complicated instructions which already is a bit worrying. Simple things should be simple to operate, but when I look at this a few things come to my mind right away. I see well, there's one display indicator for the freezer and one display indicator for the fresh food section. Let's see, letters for the freezer, numbers for the fresh food. So I figure if I change the letter, I must change the freezer temperature, and if I change the number, I should be changing the fresh food temperature, well that makes sense. And then I sort of look at the instructions and I think, well normal settings C. If I want to make the fresh food colder, I move it up from 5 to 6 or 7, okay. If I want to make the freezer colder, I move it from C to D, yeah, okay, that's sort of making sense. Well, but if I go in a little more detail, it'll get a little more complicated like I sometimes have to change both controls in order to change the temperature in the fresh food. Well okay, well I'm not quite sure what that means, so given that my model of the letters control the freezer, the numbers control the fresh food. I'm just going to stick with that as my working model. And what that might mean is that, when I look at that I have, my model could be that I have a freezer control that controls the freezer temperature, the fresh food control, controls the fresh food temperature. I probably have a cooling unit for both of them. I probably have a thermometer for both of them. And what happens is I'm controlling how much, how hard the cold air unit works, either for the freezer, or the cold air unit for the fresh food compartment works. Now, that's a pretty sensible model. It makes sense based on the controls that I could see. But suppose now I want to make the freezer warmer. Now, if I look at the instructions, there's no instructions that tell me how to make the freezer warmer. It tells me how to make the freezer colder. It tells me how to make the fresh food section warmer and colder. There's no explicit instructions for how to make the freezer warmer. So I have to rely on my model. And now I sort of think, well I guess I'll move the freezer to B because moving the freezer from C to D made it colder but what do I do with the fresh food control? Should I leave it at five of move it lower? And my model of the system, my user model really doesn't help, it really doesn't tell me. And furthermore, it turns out that the model I had of this system was wrong. The reality is, the design model was completely different and actually kind of weird. It turns out there was one cooling unit in this refrigerator. There was a valve that controls the amount of cold air directed to the freezer and fresh food compartments, that is, you have a certain amount, a single amount of cold air being produced, and you can choose to route more or less of it to each of the two compartments. You can control how cold the cooling unit is. And then there's a thermostat somewhere. And the problem is it's really difficult, or impossible, to infer this model from the system image. So, not good design. It was not easy for me to come up with the proper model of the system from what I saw of the system, the system image. Now there's a big problem too with feedback. So look at this instruction it says, after you make your changes to the controls, allow 24 hours to stabilize. So whatever change I make I have to wait 24 hours to see if it works. That's tough, that is not a design I like. Well, so we've seen some problems, so the question we have is how does a designer help users acquire the right model of a system? And the answer is, use principles of good design, and we're going to go through these principles in another video coming up soon. So that's it for the topic of conceptual models, and that's all until next time.
