Intertwining knowledge delivery and elicitation: a process model for human-computer collaboration in design

Collaboration among designers can be described with an "action-reflection-critique" model in which the explicit representation of the design contributes to a shared understanding and to the articulation of design knowledge. We describe how domain-oriented design environments based on this model support human-computer collaboration by intertwining knowledge delivery and elicitation. The KID (Knowing-In-Design) system has a shared understanding about the designers' "task at hand" through a partial design requirement specification and a solution. KID delivers design knowledge relevant to this task at hand, and the delivery helps designers uncover tacit design concerns. Designers are encouraged to store the elicited design knowledge in KID, which results in the evolution of the system's knowledgebases. The evolution affects the system's subsequent behavior by tuning the delivery toward the designers. This cycle of knowledge delivery and elicitation processes supported by KID allows designers to gradually coevolve design requirements and solutions. Acknowledgements: We thank the HCC group at the University of Colorado, who contributed to the conceptual framework and the systems discussed in this paper. We also thank Barbara Gibbons of Kitchen Connection at Boulder, Colorado, for her valuable time and her comments on our work. We thank Loren Terveen of AT&T Bell Labs, for his comments and suggestions on an earlier version of the paper. The research was supported by the National Science Foundation under grants No. 009015441 and MDR-9253425; Software Research Associates, Inc. (Tokyo); the Colorado Advanced Software Institute; US WEST Advanced Technologies; and NYNEX Science and Technology Center. Intertwining Knowledge Delivery and Elicitation: A Process Model for Human-Computer Collaboration in Design Kumiyo Nakakoji1,2 and Gerhard Fischer2 1 Practical Software Engineering Laboratory SRA Inc., 1-1-1 Hirakawa-cho, Chiyoda-ku, Tokyo 102, Japan 2Departrnent of Computer Science and Institute of Cognitive Science University of Colorado Boulder Colorado 80309-0430, USA Tel: + 1 (303) 492-3912 Fax: + 1 (303) 492-2844 [email protected], [email protected] Abstract. Collaboration among designers can be described with an "action-reflection-critique" model in which the explicit representation of the design contributes to a shared understanding and to the articulation of design knowledge. We describe how domain-oriented design environments based on this model support human-computer collaboration by intertwining knowledge delivery and elicitation. The KID (Knowing-In-Design) system has a shared understanding about the designers' "task at hand" through a partial design requirement specification and a solution. KID delivers design knowledge relevant to this task at hand, and the delivery helps designers uncover tacit design' concerns. Designers are encouraged to store the elicited design knowledge in KID, which results in the evolution of the system's knowledgebases. The evolution affects the system's subsequent behavior by tuning the delivery toward the designers. This cycle of knowledge delivery and elicitation processes supported by KlD allows designers to gradually coevolve design requirements and solutions. Collaboration among designers can be described with an "action-reflection-critique" model in which the explicit representation of the design contributes to a shared understanding and to the articulation of design knowledge. We describe how domain-oriented design environments based on this model support human-computer collaboration by intertwining knowledge delivery and elicitation. The KID (Knowing-In-Design) system has a shared understanding about the designers' "task at hand" through a partial design requirement specification and a solution. KID delivers design knowledge relevant to this task at hand, and the delivery helps designers uncover tacit design' concerns. Designers are encouraged to store the elicited design knowledge in KID, which results in the evolution of the system's knowledgebases. The evolution affects the system's subsequent behavior by tuning the delivery toward the designers. This cycle of knowledge delivery and elicitation processes supported by KlD allows designers to gradually coevolve design requirements and solutions.