[Next] [Previous] [Up] [Top]

1.0 INTRODUCTION

1.4 THESIS

The design process could be said to be fully understood if product design and manufacture were reduced to the push of a button. Unfortunately, not enough is known about the design process to automate it in such a fashion, and as the knowledge of design is expanded, unforeseen territory is often entered and the goal of automation may seem all the more elusive. In the interim between now and the day that design theory becomes a completely understood law, a system must be present to mediate among disparate domains and processes within the complete product life cycle.

Posing the computer as a medium and the design artifact as the text for communication, this paper describes the research and development of the requirements of a concurrent CAD environment. This environment is one of mediation between a number of opposing concerns, including the following:

Preliminary and formal design.
Established technology and organization of corporate information and newer forms including client/server and object technology.
Historical and future design states.
Declarative and imperative design evaluation.
Automation and assistance.
Complex model development and simple access to modeling elements.

A General Design Theory (GDT) was proposed by [Yoshikawa 81] as a guide for the construction of CAD systems. General enough to be applicable to any design domain, it attempted to scientifically formalize the relationship between man and computer, discarding the phenomenological approach to previous design theories. [Yoshikawa 83] enunciated issues of design language and of the mapping between the specifications of a function space and the solutions of an attribute space, as well as the use of production rules.

GDT attempted to objectively identify issues of importance in the human approach to design using computers, independent of the biased experience of the engineers who developed other systematic methods of design. Like this research, it did not attempt to impose another method of design, but rather implied an independence of the design artifact solutions from the functions that develop it.

This research acknowledges the role of experience in engineering design, particularly at the early stages of design. But there must be some attachment of computable structures to make this experience accessible through the computing medium. It is asserted that there are fundamental elements of design modeling which can be compounded to general structures to simultaneously address both computational and experiential aspects of design. Acknowledging that some structure is required for purposes of computer facilitation, the Analysis-Synthesis-Evaluation (ASE) model of systematic design is shown to be an effective paradigm within which to categorize and develop design agents. At issue is not whether the ASE model is descriptive or prescriptive but rather how it can effectively delineate the elements of a flexible and open CAD environment.

[Next] [Previous] [Up] [Top]