Publications / 2003 Proceedings of the 20th ISARC, Eindhoven, Holland

Automation of the Design Process

Peter John Gardner
Pages 327-332 (2003 Proceedings of the 20th ISARC, Eindhoven, Holland, ISBN 978-90-6814-574-8, ISSN 2413-5844)

Over the last 40 years ever more sophisticated computer hardware and commensurate developments in software have enabled much design to be computerised. More recently systems integration has allowed software to automatically pass data from package to package. This has effectively automated elements of the design process.

As with all developments there have been advantages and disadvantages. Automation has brought great efficiency gains and has removed many of the tedious aspects of design, but at what price? Generally designers resist the notion of automatic design and prefer to talk about efficiency of processes and the computer doing the “number-crunching”, leaving the designer “free to think”. However, despite the reluctance to acknowledge the phenomenon, [at least partial] automation of the design process has arrived. The reason why designers don't like the notion of automatic design is that automation implies loss of control and all designers should be (and are) fully responsible for all aspects of the process however efficient/automated it becomes. Whatever the semantics, there are major issues surrounding process automation and this paper explores the pros and cons in detail.

The benefits of design automation to the industry, and society at large, are considerable. Design is now faster and more accurate, and the whole process has been significantly enhanced by the available technology. However, it can be argued that the more automated the design process becomes, the more the designer loses the intrinsic feel for an appropriate solution. There are at least two documented significant structural collapses that have been, at least in part, attributed to computerised design. Lessons from both these failures are discussed in the paper. There are significant implications for the education and training of technical designers and at a more fundamental level, their basic skills-set. It is the fundamental requirement for an understanding of appropriate solutions that provides the link between automated design and the education and training of designers.

The paper does not argue that there are inherent deficiencies in computerised design but that there are differences between computerised and manual design than need be recognised, understood and managed. The effects of computerisation is so profound that the high-level numerical skills of engineering designers are now largely redundant but there is an even greater need for a deep understanding of behaviour and a "feel" for appropriate solutions. The paper concludes that the education and training of designers will have to change to reflect the new demands of the computerised design environment.

Keywords: automation, computerised design, de-skilling, education, IT, process integration, training, skills