لطفا منتظر باشید ...
20.3 The Conceptual Framework
The conceptual framework starts with the commonalties between the situated and the procedural approaches. Both approaches appear to share some minimal assumptions about human actors. First, human actors are boundedly rational and have only limited knowledge about the future. Consequently plans (as well as process maps or work flow descriptions) are often imperfect, for they typically cannot account for all possible circumstances. A process support system will therefore have to allow for run-time changes to the original plan and will have to provide contextual information about the running process to the actor as a basis for reasoning about the possible next steps. Process maps, a representation of plans, can serve as part of such contextual information (Suchman 1987; Weick 1979; Bardram 1997).Second, as Newell and Simon (1972) point out, our environment includes well-structured and less-well-structured problems. Consequently we have problems with well-defined solution strategies and others, where the solutions strategy is rather unclear (Rock et al. 1990). The transparency of the solution strategy (which can be represented as a process map) may change over time as our understanding of the problem changes. As an elusive problem becomes, for example, better understood its solution strategy may become easier to determine. Or a seemingly simple problem may become highly complex, as new facets of the problem emerge during problem solving, rendering the original solution strategy inapplicable.
20.3.1 The Specificity Frontier
The first consequence of this approach in regard to the enactment of activity is that the specificity of process structure changes over time. Bernstein and Schucan (1998), for example, provide a description of how the money-transfer process gained specificity over time. Before the formalities of banking were established, this process started as a vaguely specified process involving an ad hoc letter sent by a courier. With increasing maturity of the banking industry, the specificity of the process increased significantly. Today a money transfer is a fixed computer-based interbank clearing process with a fixed set of attributes. This illustrates the major pillar of this conceptual framework: organizational processes lie on a continuum from highly specified and routine processes at one extreme to highly unspecified and dynamic processes at the other extreme. I call this continuum the specificity frontier (see figure 20.1). A whole series of points on this frontier are possible, from a highly specific to highly unspecific.
As figure 20.1 depicts the concept of a specificity frontier in some sense bridges the gap between the structured WfMS and the unstructured communication systems. It allows for the coexistence of well-specified and almost procedurally executed processes (traditionally supported by WfMSs), and emergent situated processes (typically supported by communication support systems). It also argues that those two types of processes are at the extremes of a frontier of processes. It proposes that the whole range of processes, from highly specified and routine to highly unspecified and dynamic should be supported.
Figure 20.1: Specificity frontier
Heidi's problem, for instance, starts out as having a reasonably well-specified solution strategy (process). When she, however, realizes that the truckers in the EU are on strike, the process suddenly becomes much more problematic: the known description is not applicable anymore. Thus a support system that allows processes to start out as being well defined (and supported by a WfMS-type technology), and lets the structure become flexible (and supported by groupware technology) as soon as she finds out about the strike, would be ideal for her.
Consequently a model of business processes should be able to capture a range of process specificity (from well specified to highly unspecified). A process support system should be able to interpret process models with varying degrees of specificity. Furthermore it should support users when changing the specificity of the processes at run-time. In achieving those goals, it can close the specificity gap (pictured as a question mark in figure 20.1) between traditional process-support systems and communication support systems, and thus bridge systems following the work flow tradition and the situated action tradition.
20.3.2 Emergent Activity Relies on Structure
The second consequence of those commonalties (i.e., bounded rationality and varying specificity of tasks) is illustrated by Orlikowski (1996), who shows how change can be understood as a series of improvisational embellishments to existing practice. In other words, the actors attempt to solve the problem at hand following their interpretation of the structure and the current context. This illustrates the second pillar of the conceptual framework: that emergent activity relies on some form of structure and thus some form of specificity. Emergent activity surfaces ''unpredictably from complex social interactions''(Markus and Robey 1988, p. 588). However, we may be able to support it by supplying a fertile environment for new solutions to emerge, ''much as does a supersaturated solution in the moment it is disturbed''(Mintzberg and Waters 1985, p. 267). For example, jazz improvisation, a type of emergent activity, depends on the actors ''having absorbed a broad base of musical knowledge''(Berliner 1994, p. 492). Analogously, people in an organizational context must have some foundational knowledge about the task at hand. In addition, as Weick (1998) points out ''improvisation does not materialize out of thin air'' (p. 546). People need something to improvise on. This explains the limited success of communications support systems for business process support: from an improvisational standpoint, human actors using those systems incur the overhead of having to understand the context of the task at hand as a basis for improvisation. In the domain of organizational activity, a process map with a low degree of specificity and information about the enactment context could help actors in their sense-making, provide a basis to improvise on and thus a fertile environment for emergent processes.Consequently any system that plans to support emergent activity (which is what all activity is to some degree following the situated action approach) should provide some structure as a contextual basis for situated improvisation. Process maps (in analogy to geographical maps) can provide such structure.
20.3.3 Other Requirements
Previous research (see Swenson 1993; Abbott and Sarin 1994; Ellis and Nutt 1996; Krammer et al. 2000 among others) has shown that a process support system also should allow for the change, composition, and execution at run-time as well as provide a means of integration into an existing environment (e.g., using an open interface).
