New Developments in Engineering Design Theory (Part 1)

© T. Love 1995

A version of this paper (parts 1and 2) was delivered in 1995 as the I.Mech.E and I.E.(Aust) Annual Joint Lecture (Perth, W.A.)

Abstract

In this paper, the main developments of engineering design methodology over the last thirty years are discussed. Design methods depend on a range of design theories. By looking at the different theories it is possible to get a better idea of which design methods are best suited for particular engineering projects. A new straightforward way of looking at design and the management of design will be described and recent directions in design thinking will be indicated, especially with regard to social, environmental and ethical issues in engineering design. Sources of information for design methods and theory are provided.

Preamble

This paper is intended to give some exposure to the work being done in the ‘backroom’ of design research, a part of the discipline of engineering that receives little public attention.

First it is necessary to draw a distinction between the two terms ‘engineering theory’ and‘design theory ‘ which are often used interchangeably: ‘Engineering Design Theory’ or its abbreviated form ‘Design Theory’ has been used as a catch-all to include anything theoretical which an engineer might use. In this paper, the following distinction will be made.

Engineering Theory - is what engineers use to model situations in attempting to gain more information. It includes the mathematical models and calculations used by engineers.

Design Theory -is what design researchers use to model the activity of designing.

The difference between the two ‘theories’ reflects the difference in perspective needed to understand work in this area. Sargent 3 provocatively explains the situation as,

‘Research into design is actually a scientific not an engineering discipline, and it has to involve engineers because they are the subjects of the study and not the proponents.’

To emphasise this difference in perspective between engineering research and design research, the technological issues which engineers and engineering designers are concerned with can be contrasted with issues of design research. For example, one of the most important aspects of design research is, ‘The study of how potential ideas might be influenced before they are conceived.’ It is difficult to see how this could be conceived as a purely technological study.

A Brief History of Engineering Design Research

Research and theorymaking about the activity and processes ofdesigning is reckoned to have begun around the late 1950s (see for example; Jones (1964) 4 and Cross (1993) 5 ). The three main disciplinary focii of what was known as the Design Methods Movement were Engineering, Architecture and Planning. The following

Theorists and researchers started first with an underlying conception of the designer as ‘magician’ 6 . As a consequence of this, the focus was on unravelling the details of a problem so that a designer’s ‘magic’ could be used on it. This theory fitted well with the development of the new methods for handling complex problems which had been developed during and after the Second World War.

Almost immediately, the role of the designer in design theory changed twice. First, the new Systems perspective was not only applied to the problem, it was also applied to the designer. From this came the idea of the ‘designer as a computer’. ‘Inputs’ (including some form of ‘problem definition’) were ascribed to a system representing the designer. The designer ‘synthesised’ a solution and then ‘evaluated’ it. This process was repeated inside the ‘designer system’ until an acceptable solution was ‘output’ for others to see.

The second change in design theory was to completely remove the designer from the model of designing. Designing became a process viewed as happening in a ‘black box’ with consisting only of inputs, outputs and some unknown transformational process between them.

Removing the designer from the scene paved the way for the development of a myriad of systematic design methods and a wide variety of speculations on the best way to describe design and design process.

The most common ways of modelling designing and design process in the Engineering domain has been to use a flow chart underpinned by a view of designing as the transformation of information A typical example is illustrated in Ertas and Jones (1993) 7 for the training of American engineering designers.

The idea of the design process as a flow diagram of information transformation is so commonplace and apparently ubiquitous that the suggestion that it is based on faulty premises 8 and rarely used by engineers 9 , may come as a surprise. The search for a more appropriate basis for design theory must be directed elsewhere.

Taking a broad sweep over the literature relating to research into designing, there are almost as many theories as there are theorists. Design is most commonly seen as a process. Four contradicting views are found in which this design process is regarded as moving:

·       From an ABSTRACT statement of a problem to a CONCRETE solution.

·       From CONCRETE needs to an ABSTRACT specification of a solution.

·       From an ABSTRACT statement of needs to an ABSTRACT specification of a solution.

·       From CONCRETE needs to CONCRETE solutions.

With this level of disagreement, establishing the epistemological status of the entities of design theory becomes difficult, if not impossible.

Another, potentially more fruitful, way of classifying the different design theories is as follows:

·       Artefact based theories

·       Management based theories

·       Process based theories

·       Information transformation theories

·       Axiomatic theories

·       Theories of designing as a human action

·       Philosophically based theories

There are two main positions with regards to these theories. Some theorists see the creative activity of designing everywhere, and other see designing as something that something that occurs some times and in some places. The first point of view may be summarised as:

·       Design encompasses almost all human activity.

·       Engineering design relates to anything that is at least in part technological.

One of the main problems with such a wide definition is that it is so wide as to be almost useless - unless a strong case was to be made that ‘designing’ is a primary human activity like’thinking’ or ‘feeling’. No one, as far as I know, has been bold enough to make this their theoretical mainstay. It would require a substantial amount of justification. A further problem is that, in academic terms, this view of designing cuts across almost every other academic’s turf. All other disciplines would then be expected to addresss this issue of designing - or arrange for a design theorist to address it for them.

The other way of viewing design is to see it as a particular activity which is associated with other activities or which needs other activities to support it. A parallel may be made between this perspective on designing and the act of painting pictures. A painter, such as Van Gogh, may do a variety of activities such as; eating lunch, sweeping his studio and cleaning his brushes. These activities are necessary, and they support the act of painting. But, sweeping the floor is not painting.

There has been considerable confusion over this issue in engineering design research. Engineering designers undertake many activities which are necessary to support the act of designing. These activities may include: looking up data, filing drawings, using mathematical models, or even banking their pay cheque. The creation of an artefact may be supported by these activities, but should they be considered as part of the act of designing? (To assume this is to move towards a position that ‘everything is designing’.) The most fundamental confusion is between mathematical analysis and engineering design. Mathematical analysis (or engineering calculation) is necessary. It is the means of adding value to information. If information or data is not available directly, then often the designer may obtain it by calculation or modelling, thus it is a more sophisticated or complex way of ‘looking up data’. Regardless of its current academic and professional status, engineering calculation must be seen to lie in other realms (data collection, for example) rather than designing.

How can it be decided whether something is designing or not? The test which seems most obvious (to me at least) is to ask, ‘Is this activity exclusive to designing or is it just an example of an activity which is common to many other situations?’ To give an example: Design management is often taken to be an essential part of design, and therefore it should be researched under the discipline of design. Using the above test would indicate that design management is an example of the more general activity of managing and therefore would be best researched in the disciplines in which management is studied. Applying this test to each area of design leaves very little left! Even information processing is devolved elsewhere. What is left is what is essentially design.

What I wish to propose here, is that consequent on the above, a human centred perspective on designing is the best approach. It is not the only approach, but it does seem to provide better theoretical coherence with a wide range of other bodies of knowledge. It simplifies engineering design research as a discipline, offers a philosophically justifiable basis for its epistemology and ontology, and perhaps most importantly, fits well with the extensive work done by researchers and theorists of design over the last 40 years. A human-centred perspective towards design implies the following assumptions:

·       Design is human creative activity.

·       Design is not routine activity.

·       Only a small part of creating something is designing.

·       Other activities such as management, analysis, modelling, information gathering, financial control, and decision making are necessary, but are not designing. (These, however, in their turn,may also contain some designing.)

This may be seen as a new paradigm of engineering design, but the idea that designing is best considered as a human activity has been well established in other design disciplines for some time. From the perspective of industrial design, Jones [1] in 1970 offered an idea of the self organising designer illustrated below. As early as 1964, Alexander [2] was suggesting that the difference between primitive building and architecture was that architecture was done by human architects who design in a ‘self-conscious’ manner. (This may be seen as a similar concept to that of Schon’s ‘reflective action’.) Schon [3] developed his concept of the ‘reflective practitioner’ , in the mid 1980s. This concept although not generally taken to be central to the question of what design is, has made frequent appearances in the literature of design research in a variety of creative disciplines. From a hermeneutic perspective, Coyne [4] has discussed the necessity for including the valueladenness of the designer in any theory of designing, and technological commentators such as Crane [5] have deduced that adequate treatment of ethical issues depends on such an inclusion.

Implications of focusing on Design as a human activity

The perspective of design as a human activity resolves many issues which were problems in perspectives of design that are not human centred, and other problems are transformed. Dealing with qualitative issues is a major problems for researchers using more mechanistic theories of design. The problem is usually expressed as, ‘How can qualitative issues be expressed quantitatively so that they can be incorporated into mathematical models.’ It may be said that there is no epistemologically satisfactory solution. Using a human centred theory of design results in a reversal of the problem. Consider the following.  

1.     Most engineering information is quantitative.

2.     Designers’ creative processes including their use of ‘design worlds’ and their evaluation of partial conceptualisations are fundamentally qualitative.

3.     Mathematical models are best seen as ‘data collection’ rather than as part of designing.

Therefore, the above problem no longer exists for qualitative issues. For quantitative issues, such as the information gleaned from calculations, the issue is resolveable. The question then becomes,

‘How can quantitative data be converted to a qualitative form that is more easily useable by designers?’

Practically, examples of this sort of problem are now being seen in situations where large volumes of data are being dealt with (eg, electricity generator control rooms, satellite data analysis, music recording studio consoles and computerised design aids ). Research in these matters is normally undertaken within the provinces of psychology and human factors engineering, or as it used to be called ‘ergonomics’.

Social, Environmental and Ethical Issues

Social, environmental and ethical issues in design have both quantitative and qualitative aspects.

Quantitative - Bounds or criteria defined by government agencies, Professional Codes or Law (eg. maximum amounts of gaseous emission). These may be dealt with like any other quantitative information.

Qualitative - issues which depend on human values. As designers function qualitatively and this provides the means whereby social, environmental and ethical considerations can influence which designs are conceived and how they are evaluated [i] .

As indicated above, the problem of including social, environmental and ethical issues into a human centred theory of engineering design is straightforward. (This is not to say that its detail may not be complex or difficult, only that it has a more coherent philosophical foundation.)

Management of Design

Perhaps the most important information needed by managers is ‘Who is doing What?’ For the manager of design activities there are three aspects of designing which come directly out of the above human centered perspective on designing. These are:

·       Others, besides designers, contribute to the design of engineering artefacts and systems.

·       Most of what good engineering designers do is not ‘designing’.

·       Knowing the difference between ‘designing’ and other activities enables designers and design departments to be more efficient and cost effective.

Hubka and Eder 10 have identified the different subjects which may be associated with designing in mechanical engineering. They differentiate between subject areas in ‘general design science’ and other areas.In these other areas, the source of expertise lies in a discipline other than design.

The activities of designing and other activities associated with designing may be differentiated via a process network or time line analysis. The analysis of Ross’ 11 shows how the majority of the activities in the design process are associative activities. Interestingly some of the activities Ross regards as design, e.g. ‘do trial design’ are not necessarily exclusively design activitiesbecasue they may also include non-design activities.

Using Models for the Management of Design

Most general process models of engineering design are inaccurate at best and totally unrepresentative at worst. To be useful to a manager, a model must be a close match to the design process which is managed. The situation being modelled depends on the time, the place and the people involved in the process. Researchers involved in the management of complex systems (see for example; Flood 12 , and Flood and Jackson 13 ) have evolved methods that deal with human centred activities. These methods align well with, and may form the basis of, a human centred perspective on designing.

Using models of design processes in the management of design activities requires the manager to :

·       Construct a management model of the particular process under scrutiny - identify the different activities in detail and investigate how the activities function together.

·       Use the expertise from the disciplines that most closely fit the different activities (together with systems management methods) to improve how the process works.

·       Break down the work of the design team into components such as: information gathering, time management, financial management, evaluation, decisionmaking, mathematical modelling, communication, routine administration and creative design.

·       Look for shortcomings and redundancies in each category of resources.

·       Use this information to guide management strategy, employment strategy, investment strategy and organisational structure.

Summary

The above theoretical emphasis on human action in design results in decisions about the implementation of technology in human affairs being located (theoretically, at least) in humans. It acknowledges the role of well trained professionals using their skills, rather than reducing the role of designers, engineers and managers to that of machine minders. In this sense, it provides a representation of the activities and processes of design which is closer to reality.

To summarise the main points developed in this paper:

·       Designing is human creative activity

·       Design process includes designing along with other activities

·       The expertise in most activities in engineering design processes lies outside design (and often outside engineering).

·       Make pertinent local models of design processes (using information from Design Theory as appropriate).

·       Use local design process models to guide management strategies, investment and decisionmaking.

·       Use appropriate sources of disciplinary knowledge, ie, Design for design, Analysis for analysis, Management for management. . . .

Journals containing information about Engineering Design Research

·       Design Studies Oxford: Butterworth - Heinmann.

·       Journal of Engineering Design Abingdon: Carfax Publishing Company.

·       Research into Engineering Design New York: Springer-Verlag.

·       Artficial Intelligence for Engineering Design, Analysis and Manufacturing (AI - EDAM), Cambridge University Press, USA

References

1          Love, T 'New Developments in Engineering Design Theory: Part 1' News Bulletin Institution of Mechanical Engineers (Australian Branch), NSW. No 123 (1996)

2          Love, T 'New Developments in Engineering Design Theory: Part 2' News Bulletin Institution of Mechanical Engineers (Australian Branch), NSW. No 124 (1996)

3          Sargent, P 'Give us the tools and we'll give you doorknobs' Times Higher Education Supplement No (30.3.90) (1990) p p. 15

4          Jones, J C and Thornley, D G (eds) Conference on design methods The Macmillan Company, New York, (1964)

5          Cross, N 'Science and Design Methodology: A Review' Research in Engineering Design Vol 5 (1993) pp 63-69

6          Jones, J C Design Methods: seeds of human futures Wiley-Interscience, London, (1970)

7          Ertas, A and Jones, C J The Engineering Design Process John Wiley and Sons Inc., USA, (1993)

8          Dasgupta, S Design Theory and Computer Science Cambridge University Press, Cambridge, (1991)

9          Ullman, D G 'A Taxonomy for Mechanical Design' Research in Engineering Design Vol 3 (1992) pp 179-189

10         Hubka, V and Eder, W 'Design Knowledge: Theory in Support of Practice' Journal of Engineering Design Vol 1 No 1 (1990) pp 97-108

11         Ross, I M Effect of Organisational Procedures on Design - An Outline of the Problems. In S A Gregory (ed) The Design Method, Butterworths, London, (1966), pp 269-277

12         Flood, R L Liberating Systems Theory Plenum Press, New York, (1990)

13         Flood, R L and Jackson, M C Creative Problem Solving: Total Systems Intervention John Wiley & Sons Ltd., Chichester, UK, (1991)



[i] ‘Evaluation’ here is used in the sense of drawing out their ‘value’. It is human values which are being discussed here, not mathematical ones.



[1] Op cit

[2] Alexander, C, Notes on the Synthesis of Form, Harvard University Press, Mass, (1964)

[3] Schon, D, The Reflective Practitioner, Basic Books, New York, 1983

[4] Coyne, B, ‘Objectivity in the design process’ in Environment and Planning B: Planning and Design, Vol 18, (1990)

[5] Crane, J A, ‘The Problem of Valuation in Risk-Cost-Benefit assessment of public policies’ in E F Byrne and J C Pitt, Technological Transformation: Contextual and Conceptual Implications, Dordrecht, Kluwer Academic Publishers, (1989)