Conventional information systems provide information to actors through updating representations of states of the world. (Weber, 1997) As we have argued elsewhere (Johnston et al., 2005), this is because conventional information systems are based on the assumptions of a deliberative theory of action; that is, that an actor creates a mental model of the state of the world before acting. The conventional information system supplies information about the state of the world to inform the actor’s mental model. The situated approach is based on our insights that information about the action context and the action dependencies are both necessary and sufficient for the actor to act routinely. Situated systems also use a form of representation to provide information to the actor but in their case it is the possibility for action that is represented in order to inform the actor about satisfaction of the action dependencies. However, whereas representation is the ‘essence’ of conventional information systems (Weber, 1997), situated systems enable routine action through manipulating the action possibility space to ensure that the action context is appropriate to the required action.
We have shown that in an action abstraction hierarchy, the successful execution of an action is dependent upon the completion of the action or actions on which it is dependent. This means that in order to act, actors need to know that the actions on which their intended action is dependent have been successfully completed. Visual and/or aural information can indicate to the actor when this is so by representing the completion or result of the prerequisite action(s) or, more generally, by representing the possibility of the next action. For example, a green light may signal to an actor that they can proceed with an action.
We have discussed how the specifics of the action context mean that particular actions are constrained or enabled, depending on the affordances of the environment in which the particular time, place, actor and resources are located. Changing the affordances of the environment, through changing the physical, organisational, or temporal structures, can increase efficiency and can provide situated information to actors about what to do next and how to do it. It can also increase effectiveness by making incorrect and possibly dangerous actions impossible to perform.
For example, with regard to organisational structure, assigning a particular role to a particular class of actor circumscribes the actions available in a particular action context. In the workplace, this means that employees know that only certain types of employee can do certain things. It can also cause bottlenecks if the class of actor required to conduct a particular action is a scarce resource. Separating tasks into different time blocks is a way of structuring the temporal environment. This circumscribes the types of actions available in that time block and may increase human efficiency (Bevan, 1995) through reducing cognitive effort.
The two approaches of representing the possibility of action, and manipulating the action possibility space, can be combined to yield lightweight information systems. For example, performing a network backup may require that all users have logged off but it is inefficient for the IT technician to have to make sure that each user has logged off. Instead, a temporal structure is imposed: users are told that they need to have logged off by 6pm. The clock striking six represents completion of logging off by all users. This is a signal to the IT technician that the backup can be performed.
Hence, structures in the environment such as particular divisions of labour, physical layout, provision of tools, and timetables can enable and constrain particular actions as well as provide situated information to actors about what to do next and when and how to do it. The signals can be quite simple, even binary, and can be conveyed by the mere presence or absence of a token (icon, card, tone, light) in the environment.
We suggest that these situated systems are more compatible than conventional information systems with the human efficiency requirements of time-constrained environments. Simplifying the environment can reduce an actor’s cognitive load as well as the physical effort of searching for information. As well as being parsimonious, we suggest that these systems are more reliable than conventional systems since the potential for human error is reduced.