Towards Information Exploration Support for Engineering Designers

many heterogeneous sources of information. This is a time consuming and difficult task. This fact combined with the rapidly emerging information super highways (internet, WWW, etc.) makes it highly relevant to discuss what characterizes engineering designers information need and usage, and what kind of computer-based support of exploration activities should be provided. These are the main questions addressed. Modern engineering design often requires localization and utilization of many heterogeneous information sources. In concurrent engineering this becomes even more complicated due to the many different approaches and expertise involved. Information localization and utilization activities are complex and only insufficiently supported today. This paper reports from a case study addressing information exploration aspects of a large engineering design project. The aim was to investigate what information the design engineers used when carrying out their design work, how they used the information, which sources were accessed and how the designers get access to the information. Based on this analysis a first list of overall requirements for how information exploration can be supported by means of computer-based tools is established. Two findings were essential: The most important type of in formation concerns potential contact persons having specific expertise and knowledge, and the most frequent applied search approach was similarity search, e.g., search for a similar problem, idea, component, type of project, etc. The study, furthermore illustrated that engineering designers’ exploration must be supported through a seamless switching between many types and sources of information. Based on a case study of a concurrent engineering team of engineering designers having different background and expertise this paper identifies a set of essential characteristics of the exploration and integration of specialized information from different domains design engineers conduct during their work. This leads to a set of requirements for how these activities could be supported by means of computers. Much research in Concurrent Engineering, Engineering Design, HCI and CSCW aims at understanding what characterizes design work. Bucciarelli (1984; 1987) has illustrated the general need for thorough studies of engineering work. Others have studied what kind of information designers need in order answer questions about and verify designs (Kuffner and Ullman, 1991), which kind of information classification structures engineering designers should be provided with (Court, 1995), relevant tools and information systems platforms for engineering designers (Blessing, 1995; Govindaraj, 1997), how to provide handbook information (Burns et al, 1997), or which information sources engineers use and how they are used for finding product information (Wall, 1986). Studies of how actors in specific information domains request and search for information (Pejtersen, 1989), and establishing general requirements for how to structure electronic text information (Dillon, 1994) have also been carefully conducted. Our own previous work has explicitly addressed how engineering designers coordinate their distributed collaborative activities (Carstensen et al., 1995a; Carstensen and Sørensen, 1997). Up till now, however, most studies have had an analytical approach aiming at understanding behavior and problems. INTRODUCTION The manufacturing industry is confronted by increasing demands for improved quality products, shorter lead-times, higher flexibility, etc. One answer to this has been to apply the concept of concurrent engineering bringing knowledge of all stages of the product life-cycle—design, development, production, use and destruction—into the early phases. The result is, however, involvement of many people having different competences in the design. Concurrent engineering design thus becomes even more complex. Teams of collaborating actors have to make design decisions which have to be based upon as qualified a basis as possible. In order to do so there is a need for exploring possibilities, potentials, etc. in This is a reprint. Published in Subra Ganesan (ed.): Advances in Concurrent Engineering—CE97, Technomic, Pensylvania, USA, pp. 26-33. Most of the studies have not been explicitly directed towards actual design of support systems. tightness of control or rigor. We do not believe that one empirical effort necessarily needs to encompass both aspects. We do, however, recognize that since the results reported in this paper are drawn from a single field study, we can neither make claims as to the generality of the findings, nor to a rigorous research approach. The organizational culture at Danfoss favors that work is primarily organized in projects. It is reasonable to assume that the type of project work studied can be made subject to some generalization if an organizational culture of a similar nature is observed. First our approach is briefly introduced including a brief introduction of the work setting observed. A set of the most central findings from our field study is described, and based on these a number of requirements for computer-based support are identified and discussed. The paper is concluded with a brief discussion of the requirements and an introduction of our ongoing work. RESEARCH APPROACH The engineering design project addressed was finished when we made our investigations. It was therefore not possible to observe the actors while working. All the statements and descriptions in the following are based upon a set of retrospective comments and reflections from the involved actors and on a detailed analysis of the structure and content of the project files. This reduces, of course, the possibility of describing how concrete activities were undertaken. In order to obtain a coherent understanding of—and to design computer-based tools for—manufacturing, field studies are essential (Keyser, 1992; Siemieniuch, 1992). This paper is based on data collected in an empirical study of one development effort at Danfoss, a large Danish manufacturing company. The company employs several thousand people including large departments of both development and manufacturing. The study mainly focused on which information the engineering designers used in their design work and how they got access to this information. The field study and the preliminary data analysis was conducted over a period of ten months and was exclusively based on qualitative data collection techniques such as qualitative interviews (Patton, 1980), questionnaires and study of project files and documentation. Approximately 15 interviews were conducted by one or two of the three researchers involved in the data collection process. The work setting As mentioned, one large engineering design project was studied. The aim of the design project was to design and plan the production of the PVE II, a new and improved version of a controller for hydraulics. This is a small component to be designed into fork-lifts, trucks and similar machinery to control the hydraulically based technology. It consists of a complex electronic structure including a number of chips and flexprints encapsulated by a light-weight plastic house. It is technically a very complex component, and the description presented here is extremely simplistic. It is, however, for the purpose of this paper not considered relevant to describe the component in further detail. The research approach used in collecting and analyzing data can be characterized as qualitative research heavily inspired by both Work Analysis (Schmidt and Carstensen, 1990; Rasmussen et al., 1994) and by ethnographic approaches to studying engineering work (Bucciarelli, 1984; Bucciarelli, 1987). The data analysis was based on theories and conceptualizations for work analysis promoted in Rasmussen et al. (1994) and Schmidt and Carstensen (1990). Albeit we call it “ethnographically inspired”, our approach can be characterized as having a structured and targeted orientation. Although we did not start out with a strict set of hypotheses, we did bring an articulated perspective. We explicitly addressed information exploration needs, sources, etc. and analyzed the work according to a number of dimensions (cf. Pejtersen et al., 1995): The PVE II project lasted 4 years and there were naturally some changes in the manning of the project. Most of the time about 10 actors were involved in the project. Although they all had a background as engineers they came from rather different departments and had very different expertise, and thus different perspectives on the design to be conducted. The typical manning included the project manager and three engineers from electronic engineering. Apart from these one actor from each of the areas of mechanical design, production planning, quality assurance (QA), marketing and purchase were more or less full time allocated. Furthermore, an expert from a specialist group on environmental protection was involved as an “environmental protection guardian”. • Overall functions conducted, • The goals the actors work towards, • The constraints under which the work must be accomplished, The project manager, the electronic engineers and the mechanical engineer were physically placed in the same room, and the person from the production planning had his office close to. The others had offices in other buildings several hundred meters away. • Prototypical tasks conducted, • Types of decisions taken by the actors and strategies applied, and • The information required to accomplish the tasks. The analysis presented here is based upon the work of this group. They had, naturally a lot of interaction with other specialists, groups, departments, etc. both internally at Danfoss and with external partners. Three major external sub-suppliers were involved in the project. These were responsible for respectively: 1) Designing a complex and essential integrated circuit; 2) Compiling the electronic components and encapsuA qualitative approach offers the obvious strength of providing rich and detailed data, enabling a deep understanding of the conditions under which work is performed. It does, however, present a major limitation in terms of promoting statements of general validity. As Mason (1989) argues, the purpose of research must be to provide both the richness of detail and relevance of research problems studied, as well as a certain