Actor-Network Theory: A Literature Review

This section presents a background of ANT theory, an explanation of the SD simulation model of software evolution and a literature review of agent-based simulation modeling.
2.1 Actor-network theory:
Bruno Latour and Michel Callon described the principle of actor-network theory (ANT) in the early 1980s as a perspective for viewing complex social situations with the aim of explaining complicated interactions in a research setting [10]. Latour [9] claimed that ANT theory differs from the traditional view of social and technological theory. In the traditional view, elements forming the social situations are described as categories such as large, small, human and non-human [11, p. 3] while, according to Latour [9], ANT theory describes both human

As mentioned in chapter one, the global software process describes the collections of people and events that control software-based system evolution [3]. The purpose of developing an ANT-based model of software evolution is to reveal and illustrate better the factors under which a software-based system is evolved. By adopting ANT in the modeling of a software evolution process, it is possible to consider both human and non-human elements within the system as active elements. This enables a wider range of entities to be considered in this model than in previous models were developed by [6] and [5]. It also provides the ability to consider the software system as a participant on its own. Wernick illustrated that the first task in building an ANT model of software evolution is identifying the entities, ‘actors, mediators and intermediaries’ that make up the social and technical situation within which evolution occurs through the connections between them. Accordingly, this model is structured as 16 entities, comprising 13 actors and 3 mediators, which can be seen in Appendix

The average value is weighted against its own value from the immediate past. The ‘own weighting’ variable is contained in each participant. It refers to the percentage of the participant's own biases weighted against the other participants' degree of support with which it interacts. The degree of commitment of each participant ‘actor and mediator’ is represented numerically as the value of 1, which represents the situation in which these participants have no negative or positive attitude toward the health of the software evolution. Accordingly, the value of >1 represents a positive attitude of the participant, while the value of 1 reflects positive health, while a value of <1 reflects negative