Part III. Computer Support of Cooperative Design
“The philosophers have only interpreted
the world in different ways;
the point would be to transform it.”
Theses on Feuerbach
The chapters of Part III discuss the three major features of Hermes: the hypermedia knowledge representation, the perspectives mechanism, and the language. Hermes is an instantiation of the theory of computer support proposed in Part II. The discussion of these features of Hermes is intended to illustrate how a system based on the theory might look—a set of mechanisms for supporting the situated, perspectival, linguistic character of interpretation. While the theory suggests the usefulness of a language and a perspectives facility, many very different kinds of languages and perspectives mechanisms are possible. The particular mechanisms in Hermes that have been prototyped as part of this dissertation, suggest one possible approach. The discussion of these mechanisms should illustrate the application of the theoretical framework previously developed to the concrete design of software; these mechanisms represent an attempt to transform the philosophical interpretations into practice.
In this Part, Chapter 8 discusses the integrated hypermedia structure. This provides the medium for representing the design situation using the many media of design. The perspectives mechanism of Chapter 9 provides for flexible organization of all knowledge in the system in order to support collaboration. The language presented in Chapter 10 offers designers increased power for interpreting, communicating, and capturing their tacit understandings more explicitly.
Each of these chapters is divided into three sections. The first reviews the needs which must be addressed by the mechanisms discussed in the chapter. The second describes in some detail the implementation of the mechanisms in the Hermes prototype. The third illustrates how the explicit mechanisms are actually used by designers working in Hermes. Generally, the interfaces to these mechanisms encapsulate their computations so that they normally function behind the scenes of relatively tacit usage by designers, only becoming more explicit when the designers must articulate their understanding.
Together, the three mechanisms that are detailed here are intended to support interpretation in design. Specifically, they support the situated, perspectival, linguistic character of design. The kind of design they are meant to support is that of exploratory domains like lunar habitat design, which can be characterized as innovative in nature and collaborative in structure. The computer support proposed has been developed particularly to help designers move back and forth along the spectrum of tacit and explicit understanding. The description of each mechanism will show how it promotes tacit usage as well as facilitating more explicit understanding when that becomes temporarily necessary.
Chapter 8. REPRESENTING THE DESIGN SITUATION
Many forms of knowledge are required to support design. The lunar habitat designers in Chapter 3 used sketches of previous designs, graphical representations of design components, discussions of design rationale, terminology for thinking about the design, information from experiences of former space missions, drawings from references, and guidelines from NASA documents. They viewed problems from alternative perspectives and they deliberated issues using concepts that were redefined in the process. Rather than simply constructing a solution from these many pieces of retrieved knowledge, the designers continually modified the knowledge, trying numerous variations. They continually reinterpreted their task, candidate solutions, and the knowledge that went into the solutions.
To support what Part I of the dissertation described as the process of interpretation in design with a computer-based design environment requires a system that provides many media of representation. Furthermore, the representations of knowledge in the media must be designed to support incessant modification, tailoring, customizing, or plasticity by end-users.
According to Part II, a design environment should be people-centered, supporting the human designer’s ability to interpret and make judgments. It should support tacit usage as well as allowing designers to make knowledge successively more explicit to meet their specific interpretive needs. This suggests incorporating an end-user language for explicating terms and a perspectives structure for organizing different people’s customized versions of knowledge. To take advantage of the computational power of the computer, a design environment should provide a computationally active medium in which the designers can work individually, communicate with the computer, and collaborate with other designers on team work.
The Hermes system described in Part III attempts to meet these requirements by providing a substrate of functionality that can be used by all components of a design environment. It defines a multi-media structure in which all elements of knowledge can be defined and interconnected. All knowledge is represented as data that can be retrieved and modified by the end-user. The knowledge representation structure integrates a perspectives mechanism so that all representations of knowledge are organized into hierarchies of user-defined contexts. It also integrates a language that designers can use for defining and modifying representations of knowledge, including definitions of computer agents such as critics, queries, and displays.Section 8.1 describes the characteristics of the Hermes substrate. It discusses how it meets the requirements from the analysis of design as interpretation presented in Part I and from the theory of computer support for interpretation in design proposed in Part II. Section 8.2 shows how the substrate is defined at a more technical level. It discusses the knowledge storage, retrieval, modification, and interconnection mechanisms. Section 8.3 then illustrates how a lunar habitat design environment with multiple components can be built on top of the Hermes substrate. In addition to outlining how components for construction, rationale, specification, and catalogs can be built, it highlights the usefulness of the hypermedia, perspectives, and language in defining these components
Go to top of this page
Return to Gerry Stahl's Home Page
Send email to Gerry.Stahl@drexel.edu
This page last modified on January 05, 2004