HUMAN-COMPUTER INTERACTION
SECOND EDITION
HUMAN-COMPUTER INTERACTION
SECOND EDITION
Except when interacting with some video games, a user does not take very well to surprises. Predictability of an interactive system means that the user's knowledge of the interaction history is sufficient to determine the result of his future interaction with it. There are many degrees to which predictability can be satisfied. The knowledge can be restricted to the presently perceivable information, so that the user need not remember anything other than what is currently observable. The knowledge requirement can be increased to the limit where the user is actually forced to remember what every previous keystroke was and what every previous screen display contained (and the order of each!) in order to determine the consequences of the next input action.
Predictability of an interactive system is distinguished from deterministic behaviour of the computer system alone. Most computer systems are ultimately deterministic machines, so that given the state at any one point in time and the operation which is to be performed at that time, there is only one possible state that can result. Predictability is a user-centred concept; it is deterministic behaviour from the perspective of the user. It is not enough for the behaviour of the computer system to be determined completely from its state, as the user must be able to take advantage of the determinism.
For example, a common mathematical puzzle would be to present you with a sequence of three or more numbers and ask you what would be the next number in the sequence. The assumption in this puzzle (and one that can often be incorrect) is that there is a unique function or algorithm which produces the entire sequence of numbers and it is up you to figure it out. We know the function, but all you know are the results it provides from the first three calculations. The function is certainly deterministic; the test for you is a test of its predictability given the first three numbers in the sequence.
As another, possibly more pertinent example, imagine you have created a complex picture using a mouse-driven graphical drawing package. You leave the picture for a few days and then go back to change it around a bit. You are allowed to select certain objects for editing by positioning the mouse over the object and clicking a mouse button to highlight it. Can you tell what the set of selectable objects is? Can you determine which area of the screen belongs to which of these objects, especially if some objects overlap? Does the visual image on the screen indicate what objects form a compound object which can only be selected as a group? Predictability of selection in this example depends on how much of the history of the creation of the visual image is necessary in order for you to determine what happens when you click on the mouse button.
This notion of predictability deals with the user's ability to determine the effect of operations on the system. Another form of predictability has to do with the user's ability to know which operations can be performed. Operation visibility refers to how the availability of operations which can next be performed is shown to the user. If an operation can be performed, then there may be some perceivable indication of this to the user. This principle supports the performance in humans of recognition over recall. If not, then the user will have to remember when he can perform the operation and when he cannot. Likewise, the user should understand from the interface if an operation he might like to invoke cannot be performed.
Predictability focuses on the user's ability to determine the effect of future interactions. This assumes that the user has some mental model (see Chapter 1) of how the system behaves. Predictability says nothing about the way the user forms a model of the system's behaviour. In building up some sort of predictive model of the system's behaviour, it is important for the user to assess the consequences of previous interactions in order to formulate a model of the behaviour of the system. Synthesis, therefore, is the ability of the user to assess the effect of past operations on the current state.
Observability allows the user to evaluate the internal state of the system by means of its perceivable representation at the interface. As we described in Chapter 3, evaluation allows the user to compare the current observed state with his intention within the task--action plan, possibly leading to a plan revision. Observability can be discussed through five other principles: browsability, defaults, reachability, persistence and operation visibility. Operation visibility was covered in Section 4.3.1 in relation to predictability. The remaining four are discussed next.
In Section 8.6.1, we discussed the labelling of modes in which user actions may have different interpretations. It is generally regarded as good practice to minimize the number of modes. However, where modes are used they should be visually distinct. That is, one ought to be able to tell which mode one is in from the computer display. So, the visual presentation and the form of the dialog are intimately connected. As well as major modes, one can compare any two dialog states, and ask whether they can be distinguished from the display. If not, this may indicate a potential trouble spot where the user can get confused. These are both visibility requirements, and similar issues of observability and predictability will arise in Chapter 9.
processed in 0.003 seconds
|
| |
|
HCI Book 3rd Edition || old HCI 2e home page || search
|
|
| feedback to feedback@hcibook.com | hosted by hiraeth mixed media |
|
| |