Behaviour Models There are a number of models that predict the way in which an interface or user will behave.

Predictive Modelling Predictive modelling is a way of pre-empting what will happen without having to carry out research and delay the introduction of the interface while lots of people test it out. A predictive model is an equation or calculation used to forecast an event. The reaction time to respond to a command will vary depending on the user. The interface may be responsive to respond time and shut down/hibernate if a command takes too long

Key Stroke level model (KLM) KLM recognises very low level actions. The model breaks down each sequence of operations into individual actions such as: hitting keys on the keyboard clicking on the mouse pointing the mouse moving between using the mouse and the keyboard again. Each action is assigned a time in order to calculate how the system will respond.

The (TP) Throughput TP relates to the productivity of the computer. Throughput measures include the amount or speed of processing in response to a command. Speed processing – including variation in tasks or complexity is known as the response time -

Fitts’ Law Fitts’ law is a method for calculating throughput in advance for any system design by predicting humans movement and motion based on time and distance, called psychomotor behaviour. It was developed in 1954 to counteract the assumption that the time taken to travel from A to B is based on the distance between two points.

Fitts’ law identified that time depends upon the size of the object to be moved and the size of the object with which to move it. User time will vary according to the user, the location of an icon, menu or GUI, the click of the mouse or hitting a key and even the pressure applied.

Descriptive modelling Key action model (KAM) Buxton's three state model Guiard’s model

Key action model ( KAM ) KAM identifies the need to evaluate how the user will expect the computer to behave or react and how this may be different from how the computer actually reacts to commands. A new user may find it difficult to log on using a password and username (they don’t understand that the computer is visiting a database to verify, hence it has to be 100% accurate. A user may get frustrated when they think they have selected “Shift F” but the file menu appears because they have accidentally hit “Alt F”

Other feedback examples include the depression and suppression of buttons on dialog boxes and invitations or steering to the next most likely command by highlighting a button. Note the highlight around this button

Buxton’s three state model This model is concerned with the pressure and dexterity with which users make movements using mice and touchpad's. The interface will need to be responsive whether the command is via the mouse, a touchpad on a laptop or a button in the keyboard. A user will have their own preference as to what they are used to using so the system but be equally responsive to ensure usability.

Guiard’s Model Guiard’s model of bimanual skill relates to the preferred method of interacting with computers and input devices. When designing your interface you must consider the ease of use and you must not rely on the user always inputting data or mouse actions using a preferred hand.

The positioning on the interface must also be accessible and logically laid out if used by a left handed person. Guiard’s model identifies that users with two hands are unlikely to be able carry out similar actions to the same degree of effectiveness with either hand.

Task Research the aforementioned behaviour models, taking notes, to consolidate your own understanding.