The review of the different mechanisms for providing for the mutability of semizoic artifacts leads us to some conclusions as to how design theories for ISDT should be specified. Walls et al. (1992) omitted Dubin’s (1978) concept of ‘system states’ in their formulation of ISDT components. In Dubin’s terms, one should specify in a theory what states of a system will be covered by a theory that is proposed (see Gregor and Jones, 2004, 2006, for further elaboration). Our discussion above demonstrates that such a component is valuable in ISDT because they deal with changing semizoic artifacts, which will almost certainly demonstrate some form of mutability over their life and thus changes in state. We argue that an ISDT will be improved if the proponents of a theory consciously reflect on the degree of change they anticipate for their designed artifacts. Some degree of change may be provided for deliberately in the meta-requirements (as in being flexible in the face of future amendments). In addition, however, the theorist should document what changes in state the theory will cover.
A further interesting conclusion can be drawn by careful study of the nature of the changes that can occur in achieving the meta-requirements listed in the previous section. It is not only system states that can change but also the basic structure of the system itself. One way of conceptualising these broad directions in which information systems change is to think of an IS schema or model (structure and functions of the system) in addition to the IS states that the system can occupy at different times.[7] When thinking of the way in which an information system changes, we can think of changes both to:
its model/schema (its basic form and functional capacities), and
its state (i.e. the changes as it moves from one state to another over time).
A system’s model/schema (its basic form) is related to its design and is the subject of design theory in IS, but it is also significant to recognise state changes. A system’s capability to change its structure (IS schema) requires that the system has a reflective capability, including a self-representation (Maes, 1987). The system has a model of itself (schema) and it has the capacity to change itself by its computation (Maes, 1987).
Table I summarises our findings by showing the types of mutability discussed and their appearance at the levels of the IS schema and IS state. With some types of mutability the changes to the schema (structure) can be anticipated to some extent and facilitated (as with re-design), or the degree of change can be limited (as with nilpotent systems, homeostasis and autopoiesis). With other types of mutability the extent or nature of change cannot easily be anticipated, although the designer may set up conditions for change to take place (learning, evolution and emergence), meaning that the resultant system is only in part designed.