Paul Hendriks

Below you will find `summaries' of two papers related to collaborative spatial decision making that I am currently working on (the first is more directly related to CSDM, the other somewhat more remotely). What I did was rip some sentences out of the current versions of both papers that in combination have some similarity with a summary. Work on both is at a stage that I expect to be able to send them to a journal before the end of September.

I start with a short description of a CSDM computer lab (called the VISA skills lab) that will be operative at the University of Nijmegen starting next September. My personal connection with that lab is that I am one of its initiators, and that I am on what we call the `sound-board group' whose task it is to select, initiate, stimulate, and evaluate projects for the lab (I guess it is called `sound-board' since we make a lot of noise in the process).

The VISA skills lab

Group decision technology enhanced with visualization tools

The Faculty of Policy Sciences of the University of Nijmegen will start operating its so called VISA skills lab in September 1996. The mission of this laboratory is to offer a computer equipped environment for research and teaching centered around the integrated concepts of cooperation and visualization (`VISA' is an acronym combining the first letters of the Dutch terms for these two concepts). The laboratory is conceived as a group decision room, or `war room', with dedicated software to support group decision making (for instance the Group Systems software of the University of Arizona), enhanced with a variety of tools aimed at visualization. The laboratory is unique in several respects. In the first place this uniqueness derives from the fact that the laboratory is set in an environment of a variety of research and teaching groups within the Faculty, which comprises a Business School, a School for Environmental Studies, and a School for Political Sciences. In a number of research and teaching programs of these schools theories regarding group processes and visualization play an important role. That is: the decision room is not set in an environment of computer scientists and developers, but in an environment where theoretical and practical work on group processes and visualization flourishes. In the second place a distinctive element in the conception on which the laboratory is based is the combination of group decision approach with theories concerning visualization. This combination is important since it is essential for the support of group processes that the facilitator of these processes arrives at a generally recognized depiction of the shared environment in which the group members see what their common task is. A third aspect of the uniqueness of the lab lies in the fact that the laboratory can serve as a `market' that intends to generate synergetic effects by joining the various specialist groups. It is specifically this third element that is relevant for the field of "computer supported collaborative spatial decision making, since an explicit goal of the VISA lab is to bring researchers from the environmental group (geographers, environmental scientists, physical planners) and of the business school (organizational scientists, information scientists) together. Right now several research and teaching initiatives are being prepared that will provide the focus for these cooperative efforts.

From GIS to spatial group decision support systems: A cybernetic itinerary

Paul Hendriks
Dirk Vriens

Although GIS are used as tools to help solve problems with spatial dimensions, they are usually not called spatial decision support systems. And although being used by single persons is rare for GIS, as spatial decision making is usually a group process, they are certainly not called spatial group decision support systems. So, obviously GIS go some way as support tools for groups dealing with spatial problems, but they do not go all the way. What, we may then ask, discerns a GIS from a spatial group decision support system (SGDSS)? The answer to this question could be deduced by looking at the standard definition of the latter, and by then identifying what extensions in the first are called for.

The standard definition of an SGDSS refers to the general class of decision support systems, and identifies discriminatory features of these. The nature of decision problems, the analysis capabilities of the computer system, and the ease of use of the user interface are usually prominent elements in this definition. Next the `spatial' and `cooperation' elements are added, both in the description of the class of problems and in ways to support the solution of these problems. This eventually leads to concentrating on the selection, implementation, and evaluation of the appropriate tools. So, eventually, if these tools would be added to GIS, these systems might deserve the label `collaborative spatial decision support systems'. This approach can be characterized as `piecemeal': it starts by identifying potentially useful elements of the SGDSS, and next tries to combine them into a workable whole.

Several drawbacks are inherent here. The approach inevitably leads, for instance, to a selection process of a haphazard nature. When criteria for inclusion or prioritization are called for, these will always have to be added as a separate operation. Also: the identified set of elements will always raise doubts as to its completeness, and if essential elements have not been overlooked. An interesting alternative is offered when working the other way around, that is by aiming at a description of the `whole' and fitting useful elements for support (the `tools') in there. The paper explores this second line, and uses the conceptions of systems theory as a departure point and as guiding principles. Aim of the overall operation is to describe the concept of a spatial decision system, and to infer the prospects of supporting decisions from this description. Computerized support systems can be fit in there since these systems can be seen as elements (or subsystems) to facilitate processes in one of the major aspect systems of the overall decision system, the information system. In other words: conceiving GIS as an SGDSS can be understood within the operation of shifting the boundaries of the decision system. Core concept in describing a spatial decision system is complexity: the description of the spatial decision system is satisfactory as far as this description does justice to all identifiable sources of complexity. And also: identification of the role of automated decision support systems is a direct function of how they allow to deal with one or more sources of complexity. Four sources of complexity are described that in combination cover the ground of what a spatial decision system is: instrumental complexity, dynamic complexity, social complexity, and constructive complexity.

A valuable feature of this approach, as elaborated in the paper, is that the need to support cooperation is integrated in a natural fashion in the overall conception of a spatial decision system, since it directly follows from the third form of complexity: social complexity. That is, the cooperative nature of a decision support system is not seen as an extension of the standard conception of such a system, but as an essential part of it.

Supporting spatial data coordination in decentralized environments

Paul Hendriks

In organization theory as well as in organizational practice an unmistakable tendency is shown away from functional separation and toward task integration. As a result an increasing number of companies, and even government bodies, are organized around production processes in combination with a client (instead of product) orientation. These trends imply an encouragement of decentralization and the emergence of what is called 'integral management': cutting through hierarchies and putting responsibility as close to the level where the decisions involved have to be effected. The underlying push toward work process orientation and away from a separation of deciding and doing is at the core of modern socio-technique. This design methodology forms a central theme for research at the Nijmegen Business School. Following the ideas of this methodology, organizations are no longer divided into departments based on functional criteria (for instance, separate departments for buying, production, sales and services), but are set up around so called `complete task groups. Various other ideas and design methodologies also contribute to these developments, such as lean production, Business Process Redesign, recent conceptions regarding human resource management. In other words: both organization theory and practice show a clear push toward decentralization. The use of information systems, on the other hand, may tend to reintroduce centralization through the back door. Main reasons are synergy effects, high investment costs, calls for company-wide management reporting, data efficiency, possible exchange of data files. For technology where high investments may be necessary and large bodies of data are at stake, such as GIS, this is a fortiori the case.

A coordinated use of shared technologies between the various groups within the organization seems to offer the best way to avoid a clash between a decentralized, process-oriented overall setting of the organization, and centralization in the organization of tasks around information technology (IT), or, more specifically: GIS. In the paper specific attention is given to a central element for shared GIS use, the aspect of shared use of data resources (cf. the concept of data warehousing). Recently the subject of geographic data sharing has drawn considerable attention, particularly initiated by and centered around a research initiative of the National Center for Geographic Information and Research (Onsrud & Rushton, 1995). Although the focus there is primarily on issues of data sharing between organizations, for the most part the discussions are also applicable to intra-organization situations where multiple parties have common data needs.

Data sharing within the organization calls for coordination: a wide range of issues has to be agreed upon, including selection of data to be acquired, definition of views, division of cost, assuring updates, etcetera. Thus, what appears to be a solution, data sharing, immediately turns into a new problem: coordination. The solution of this problem reintroduces the risk of centralization. Traditionally the need for coordination is translated into the function of a coordinator (a person or group). Alternatively, coordination should not be based on central direction, but should be conducted in a participatory fashion. This presents the participators in the coordination process, the actual data users, with a class of problems with which they are usually not supposed to deal. The paper specifically focuses on how participatory coordination of spatial data sharing can be supported by computers. The paper explores how groupware may be used to facilitate an approach to coordination that avoids reintroducing centralization tendencies. Some field studies conducted at local government bodies using GIS is used to support the case for participatory, not centrally directed coordination of spatial data use.

Paul Hendriks