In spite of the advances of current CAD systems, the conceptual phases of design still suffer from lack of appropriate support tools. Contemporary research toward design support and automation puts substantial effort into the development of improved Computer-Aided Conceptual Design (CACD) and Computer-Aided Industrial Design (CAID) systems, both by industrial R&D and by science academia. However, there are indications that some of the persistent problems can only be solved if we more deeply understand what the requirements for the tools are. We extracted these requirements from two sources, from practicing designers in industry based on an inquiry, and from data generated in experimental design processes.
The inquiry indicated that improved CAD tools for conceptual design were lacking and would be welcomed if they were available. It was also explicitly revealed that even in the earliest phase of conceptual design, the enforcing of constraints to the designed shape was rated as crucial, besides an unlimited freedom of shape modeling as, e.g. offered by virtual claying. In general, a computer-based, early evaluation and analysis of design alternatives was ranked as the most urgent type of tool that should be developed.
Analysis of the protocol data from the design experiments revealed two items. First, the technical quality of the designs obtained by the test subjects suffered from a lack of intermediate evaluation. Second, the subjects often neglected to utilize an evaluation tool when it was provided.
From the extracted requirements we were able to derive recommendations for the research into better CACD tools. We also present some recent results that we obtained when we followed the recommendations.