Designing for Social Configurations: Pattern Languages to Inform the Design of Ubiquitous Computing

ly, Alexander’s creation of a pattern language can be framed as an instance of research, where an initial interest, in this case the quality of life in designed spaces, leads to a theory in a bottom-up process. From this perspective, Alexander’s work appears similar to the methodological considerations by Glaser and Strauss (1965, 1967), two sociologist, who, when confronted with the task of studying the social aspects of dying and death in hospitals, coined ‘grounded theory’ as a method for developing theories that hold relevance not only for scientists but also for practitioners. In grounded theory, the lens of research is not defined a priori; instead, it evolves in the course of the study, grounded in and targeted for the phenomenon at hand. Instead of testing predefined hypotheses based on existing theories, Glaser and Strauss look for patterns in ethnographic data that they collect through observations and interviews. For their study of dying patients they argue that “discernable patterns of interaction occur predictably, or at least non-fortuitously, during the process of hospitalized dying, and that explicit knowledge of these patterns would help the medical staff in its care of dying patients” (1965, p. ix). A grounded theory comprises the elements of categories, properties and hypotheses. Glaser and Strauss (1967, p. 36) define a category as a conceptual element of the theory that describes an abstract concept that explains a certain type of behavior. A property describes an aspect or element of a category. A hypothesis, finally, suggests a relation between the different categories. Practically, grounded theory methods can be summarized as follows. In a number of iterations, the researcher collects empirical data, “theoretical sampling” as Glaser and Strauss (1967, p. 45) named the process. The data resulting from theoretical sampling is first coded and then analyzed using the “constant comparative method” (1967, p. 101). The constant comparative method is described as a four step process (pp. 105-113). In step one, “comparing incidents applicable to each category”, the researcher identifies incidents in the data, compares them and builds categories. In step two, “integrating categories and their properties”, the focus shifts from the comparison of individual incidents to the comparison of incidents with the already existing categories and properties. In step three, “delimiting the theory”, the researcher aims at increasing the integrity of the theory. Finally, in step four, “writing theory”, the researcher, convinced of the saturation of the theoretical framework, writes up the theory. Strauss and Corbin (1990, p. 26) call for “theoretical sensitivity”, a personal quality of the researcher required to create “awareness of the subtleties of meaning of data” (Strauss Figure 1. Alexander’s concepts of structure preserving transformations and pattern languages transposed to action research and grounded theory.