Lakhasly

Design strategies
The act of making an architectural decision can perhaps be stripped
of its mystique, while some far more viable set of operations is seen
to add up to something – not a style, not even a discipline, but some
indefinable aggregate of operations which have been intelligent and
appropriate and have given a situation its fourth dimension.
Peter Cook, Architecture Action and Plan
I would be the voyeur of myself. This strategy I employed for the rest
of my captivity. I allowed myself to do and be and say and think and
feel all the things that were in me, but at the same time could stand
outside observing and attempting to understand.
Brian Keenan, An Evil Cradling
Theory and practice
In the last chapter we saw that it is common for designers to carry
some set of guiding principles with them through their working
lives. This intellectual baggage is most frequently gathered during
that career, with each project contributing to it in some way. We
saw some examples of sets of guiding principles and many others
could have been presented. The intention was simply to suggest
that it is not necessary to include revolutionary or fringe ideas about
design in order to find considerable variation in approach to the
design process. This hopefully acts as a counterbalance to the
earlier part of the book when emphasis was laid on the more theoretical
writings of design methodologists. If we are to gain any real
insight into the complexities of the design process then we must
study not only what theoreticians say but also what practitioners do.
The early years of the design methodology movement were
characterised by a tendency to look for common features in the
design process or at least to classify design strategies. Earlier in this
book we examined some maps of the design process which it is
assumed will be taken up by all designers. The message from the practitioners is rather different. They speak less of clearly defined
routes and rather more of their own individual interests, approaches
and strategies. Our earlier examination of some maps of the design
process suggested that, whilst many seemed quite logical, none
were really all that useful. The writings of practitioners confirm
the view that there is not one route through the design process
but many. However, it is not enough to rely entirely on designers’
accounts of what they do. If we could accurately describe what
goes on in our head when we design, then there would be no need
for any books such as this!
Begin at the beginning
We know that the process starts with some sort of problem and
finishes with some sort of solution, but how do designers get from
the first to the second? We have explored maps of the design
process and generally found them wanting, since they are neither
accurate nor helpful. So just how do designers begin their work?
We know that design problems are rarely, if ever, fully described
at the start of the design process. We have also seen empirical evidence
suggesting that designers use what we might call solution
rather than problem-focused strategies. That is to say their emphasis
is more on reaching a solution rather than on understanding
the problem. Our examination of the nature of design problems
and solutions perhaps now shows this to be rather more logical
than it might have at first seemed. We saw that design problems
cannot be comprehensively formulated and that solutions cannot
be logically derived from them. However, most design problems
are also far too complex for the designer to hold all the factors
in mind at once. So where do designers begin and what sort of
strategies do they employ to proceed?
The brief
Conventionally a design begins with a brief, which we may imagine
a designer is given by a client. However, since design problems
cannot be comprehensively stated this begs the question of
what is in the brief and what is not! This itself can vary considerably.
The brief may be quite complete in a design competition. In,
for example, architectural competitions there may be a site, a schedule of accommodation and a set of requirements all laid out
quite explicitly. This is necessary in the case of the competition
where the designer is probably allowed little or no contact with the
client before submission. In the more normal design process our
question is not so easy to answer. A common complaint from
designers is that their clients do not involve them early enough in
the process. Perhaps clients feel that they must have a clear definition
of the problem before they commission a designer, but this is
not so. In a study of architects and their clients, most of the architects
argued that they preferred to be involved with the project
from the very beginning (Lawson and Pilling 1996).
Some clients are experienced at their job, and may even be acting
in that capacity professionally. It is also increasingly the case that large
clients for buildings may take on their own architects to help them
develop a brief which may later be given to quite different architects.
However, many clients for design are less experienced at preparing
design briefs. The architect and interior designer Eva Jiricna tells how,
in her experience, ‘we never, ever get a brief from a client which we
can start working on’ (Lawson 1994). Now this might seem problematic
for designers, but when asked about this most of them are quite
happy to receive briefs which are very brief indeed! The Malaysian
architect Ken Yeang, even prefers to start with what might be called a
‘mission statement’ of just a few sentences. (Lawson 1994). The view
expressed by Michael Wilford describing his work with James Stirling,
is reflected by many architects and designers:
We have found over the years that the ideal brief is probably one or
two pages even for the most complex project. Many clients think they
have got to produce something that is two inches thick before an architect
can even put pen to paper. We prefer it the other way round, we
prefer the thinnest possible information so that we can get a grasp on
the whole thing and gradually embellish it with detail later.
(Lawson 1994)
Protocol studies
To find out how the design process actually begins to develop the
brief and formulate a solution we need to turn to some of the
many studies on design process protocols. These protocols have
been gathered under a wide range of conditions, but all share
in common a rather more controlled environment than the design
studio normally provides. The process studied usually has quite a
short duration measured in a few hours and often is completed within one session. Such conditions are, of course, highly artificial
so we must be careful how we analyse the findings of such studies.
Not surprisingly, most design strategies seem to begin with a
brief scanning of the problem as it appears initially. However, it
is also common to find that elements of solutions rather than
problems begin to emerge very early on in the process. In one of
the earliest of these studies, subjects were asked to design a new
bathroom, and they invariably began drawing solutions almost
immediately (Eastman 1970). One experimental technique used to
externalise and reveal design thinking is to use groups of subjects
and record their conversations. One such study of architectural
students designing a nursery school was video-recorded and then
analysed for both words and actions. It was rarely very long in
these protocols before the subjects began to use such phrases as
‘this suggests’ or ‘we could try’. It was found here that different
aspects of the problem were examined to see what they might
suggest in terms of ideas about the solution, rather than analysed
in their own right (Agabani 1980).
There are many ways of analysing the data from such design
process protocols. A notable contribution to the field has been
made by a conference at which all the contributors had analysed
the same two video-recorded design protocols. Both were industrial
design problems, in one case tackled by an individual who was
asked to think aloud and in the other case was worked on by
a group (Cross, Christiaans et al. 1996). Some researchers tried to
break down the process into sequences, others looked to classify
the kinds of cognitive activity they thought to be revealed. Others
still tried to link the events on the path to the solution with phases
of thinking, while yet others concentrated on the cognitive style
of the designers. Finally, researchers concentrated on the inadequacies
of the protocols themselves to properly represent real
design activity (Lloyd, Lawson et al. 1995). Thus there was sufficient
material here to publish a book larger than this one just on two
design protocols!
Heuristic strategies
An examination of protocols obtained from such closely observed
design sessions reveals that most designers adopt strategies which
are heuristic in nature. The essence of this approach is that it is
simultaneously educational and solution seeking. Heuristic strategies do not so much rely upon theoretical first principles as on experience
and rules of thumb.
To illustrate this principle let us look at two methods of sizing
timber floor joists. In the first, theoretical method, calculations are
performed using the known compressive and bending stress capabilities
and elasticity of the timber. The calculations give a depth
of timber which will not deflect more than 0.003 of the span and
will not cause the bending and sheer stresses to exceed the permitted
levels. The calculations are based on established theories of
structural mechanics and would be performed by structural engineers
and required for building regulation approval. The alternative
to this precise but laborious procedure is to use our second,
rule of thumb or heuristic, method. There are many possible rules
such as ‘the depth of 50 mm wide joists at 400 mm centres is
25 mm for every half metre of span’. Such a method is by no
means precise but will never be very far out. However, not only
does the method go straight to the solution, but it is educational in
the sense of clearly identifying the critical relationship between
depth and span of the joist. The rule of thumb is also much more
practical in that timber does not come in an infinite range of
depths but is commonly available in multiples of 25 mm.
This rule of thumb provides a good model of the heuristic strategy
so commonly employed by designers. A rough idea is quickly developed
for the most significant elements of the solution which can then
be checked by more precise methods and adjusted as necessary.
Such rules as those relating depth and span clearly cover the critical
aspect of the problem of sizing a joist. However, in more complex
design situations it is by no means so easy to decide what is critical.
Indeed what is important or critical is likely to be a matter of opinion.
Here designers need rather more sophisticated heuristic strategies.
Three early phases of working on the same
problem
To see how this might actually work in practice we shall briefly consider
the approach taken by three groups of architecture students
towards a competition to design a large new county authority office
building. After a fairly short period of work the groups presented
their ideas and thoughts so far. Here, then, rather than working on
protocols we can analyse the presentations made by the design
students at an early interim criticism session with their tutors. The first group started by describing how they felt that the
environmental requirements of the office space were the critical
factors (Fig. 11.1). They had done a literature review of all the
research they could find on office space and had arrived at a
sketch design of a ‘typical bay’ showing the structural and
service systems for providing shelter, power, comfort and light
while maintaining a relatively uninterrupted floor space to give
flexibility of layout. The building, they thought, could be
assembled by replicating these bays as desired and as the site
permitted.
By contrast the second group took the view that office space
itself was not difficult to design and they had focused their attention
on some rather unusual features of the site. (Fig. 11.2) The suburban
parkland site was located between two major radial roads
connected by a footpath. This group had noticed that the competition
brief had stressed the importance of not presenting a remote
or forbidding image to the ratepayers. They decided to build their
office around a covered mall which followed the line of the
footpath and thus brought the public right through the building.
Taken together with the banks of trees, south-facing slope and
considerations of screening noise from the busy roads this enabled
our second group to develop proposals for the siting and massing
of their building. The next phase, they explained, would be to fit
the various departments into the building adjusting the envelope
where necessary. The third group had focused more on the visitors rather than just
the regular inhabitants of the building (Fig. 11.3). This group were
anxious to avoid what they saw as the usual failings of such buildings,
that is, presenting large inscrutable façades with unclearly
structured interiors in which it is easy to get lost. For them the
whole structure of the organisation provided the stimulus to building
form. Each section and department were to be clearly articulated
using a hierarchy of open spaces linked by well-defined routes to a
central entrance court.
It is difficult to decide whether any of these approaches are
better than the others and it is certainly not possible to declare any
to be either right or wrong. Although at first sight these three
approaches may seem rather different, in fact they share basically
the same overall strategy. In each case a group of sub-elements of
the overall problem have been clustered together and elevated to
the role of form generator.
What differentiates the three is simply the kind of constraint
which has been used in this focal role. The first and last group
concentrated on the way the building should be organised by
focusing on internal constraints while the second group looked at
the external constraints imposed by the site. The first and second
groups looked at constraints generated by two different types of user, the employee and the local taxpayer. The first group gave
priority to the efficient control of the working conditions and thus
recognised mainly radical constraints. By contrast, the second
group thought that the quality of the place was more important
and they recognised more symbolic constraints. The third group,
when questioned, saw no conflict between these and felt that the
physical expression of the organisation achieved in their building
would not only be easy for the taxpayer to relate to but would also
lend a sense of identity and belonging to the employees, thus
creating a good social working environment.
The primary generator
We have seen how the range of possibilities can be restricted by
initially focusing attention on a limited selection of constraints and
moving quickly towards some ideas about the solution. In essence
this is the ‘primary generator’ idea which we first introduced in
Chapter 3, but where does the primary generator come from and
how does it work?
Obviously it is highly desirable that the primary generator
involves issues likely to be central or critical to the problem.
However, what is central and what is critical may turn out to be two quite different things as we shall see. The student architects
designing a building for a county administrative authority used
a variety of generators relating to the radical functions, user constraints
and external constraints of the site. The first and obvious
source of a primary generator, then, is the problem itself. Finding
those issues most likely to be central is a matter of common sense
and some experience, and these students were all demonstrating
a growing sense of judgement in these matters.
What is used as a primary generator is also likely to vary to some
extent between the different design fields and problems. Mario
Bellini the designer of the Olivetti golf-ball portable typewriter,
emphasises the difference between designing static artefacts such
as furniture, and mechanical or electrical goods in this respect
(Bellini 1977). Obviously, the product designer must learn to adapt
the design process to the situation.
We have seen in the last chapter that designers develop their
own sets of guiding principles and these often set the direction for
the primary generator in any one design project. Thus the architect/
engineer Santiago Calatrava with his guiding principles of dynamic
equilibrium is likely to use practical constraints about the structure
of his building. However, he has himself noted that this is not
enough, and that it is the highly specific and local external constraints
which often help him to create form:
I can no longer design just a pillar or an arch, you need a very precise
problem, you need a place.
(Lawson 1994)
For the experienced designer, then, the guiding principles when set
against the local external constraints may often create the material
for the collection of issues which primarily generate the form of the
solution. The designer uses this initial attempt at the solution gradually
to bring in other considerations, perhaps of a more minor or
peripheral nature.
The central idea
These primary generators, however, often do much more than simply
get the design process started. Good design often seems to have
only a very few major dominating ideas which structure the scheme
and around which the minor considerations are organised. Sometimes
they can be reduced to only one main idea known to designers
by many names but most often called the ‘concept’ or ‘parti’. In 1994 Jonathan Miller made his Covent Garden debut as an
opera director, having also designed the sets. In the programme
he wrote that ‘the formal artificiality of the work is part of its essential
mechanism, for it demonstrates reality without slavishly representing
it. It is an argument as opposed to a report – an epigram
rather than a memo’. His production of Così fan tutte was set in
modern times and relied upon costumes exclusively designed by
Giorgio Armani. The public is well used to Armani’s own restricted
palette of plain-coloured fabrics in soft textures and colours largely
restricted to fawns, beiges and browns. This simple idea was
carried through into the colours and textures of the set, itself very
simply arranged using a large backdrop wall with an opening
surrounded by a suggestion of a classical architrave. With all the
technical and financial power of the Royal Opera behind him,
Miller chose this simple and consistent message which effectively
conveyed his interpretation of ‘demonstrates reality without slavishly
representing it’. It was surely the determination with which he
resisted any temptation to depart from this one simple single idea
which made this production so memorable visually.
The industrial designer James Dyson is famous for a number of
innovative domestic products and is perhaps most well known for
his revolutionary ‘Ballbarrow’. Dyson had experience of using a
traditional barrow and found it frequently got stuck in the muddy
ground of a garden (Fig. 11.4). He transferred the idea of using a
spherical wheel from some previous experience and adapted the
shape of the body of the barrow to make it more suitable for mixing
cement and for tipping. As Roy (1993) says, throughout the
design process was ‘an essential generating idea . . . a ball-shaped
wheel’. Roy documents this and other cases where the whole
design process is driven by one single, relatively simple, but revolutionary
idea.
Another dramatic example of this is reported by Nigel and Anita
Cross in a fascinating study of the successful racing-car designer
Gordon Murray. It was Murray, when working for the Brabham formula
one team, who first introduced the idea of refuelling pit stops
since adopted by all his competitors. Murray describes how he was
thinking logically how to make the car lighter in order to make it
faster. The idea of running with a half empty fuel tank became the
central driving force behind a huge development programme. At
that time pit stops were only used in emergencies and to change
tyres. Murray worked out the gains in time from the lighter load and
calculated the maximum time he could allow for refuelling whilst still
gaining an advantage. From this came the need to design a way of injecting the fuel much faster and a way of heating up the new tyres
to racing temperature before fitting them. Both have become common
and accepted practice.
These examples from very different design fields all offer very
good examples of the creative process studied in Chapter 9.
A moment of inspiration leading to a central or big idea combined
with dogged determination and single-mindedness. Gordon
Murray’s own description of the pleasure he gets from his job
reveals this process:
That’s what is great about race car design, because even though you’ve
had the big idea – the ‘light bulb’ thing, which is fun – the real fun is
actually taking these individual things, that nobody’s ever done before,
and in no time at all try and think of a way of designing them. And not
only think of a way of doing them, but drawing the bits, having them
made and testing them.
(Cross 1996b)
This central generative idea may become very important to the
designer for whom it sometimes becomes like a ‘holy grail’.
Characteristically designers become committed to, and work for, the ‘central idea’. The architect Ian Ritchie explains the importance
of this to the whole process:
Unless there is enough power and energy in this generative concept,
you will actually not produce a very good result, because there is this
three years or so of hard work to go through and the only sustenance,
apart from the bonhomie of the people involved, is the quality of this
idea, that is the food. It’s the thing that nourishes, that keeps you, you
know every time you get bored or fed up or whatever, you can go back
and get an injection from it, and the strength of that idea is fundamental.
It has to carry an enormous amount of energy.
(Lawson 1994b)
Just as a commitment to the idea can be seen to ‘nourish’ the
designer, as Ritchie puts it, so can the search for it in the first place.
The central idea does not always appear easily and the search
for it may be quite extensive. The architect Richard MacCormac
describes this search:
This is not a sensible way of earning a living, it’s completely insane,
there has to be this big thing that you’re confident you’re going to find,
you don’t know what it is you’re looking for and you hang on.
(Lawson 1994b)
The central idea may not always be understood immediately it
begins to appear. Richard MacCormac has described this in the
development of the design for his acclaimed chapel at Fitzwilliam
College in Cambridge. (Fig. 11.5) Very early in the design process
the idea was established of the worship space being a round object
at the first floor in a square enclosure: ‘At some stage the thing
became round, I can’t quite remember how.’ Eventually the upper
floor began to float free of the structure supporting it. However,
it was not until the design team were considering such detailed
problems as the resolution of balcony and staircase handrails that the team finally understood the idea and made explicit the notion
of the congregational space being a ‘vessel’ (Fig. 11.6). This was
then to work its way right through to inform the detailing of the
constructional junctions which articulate the upper floor as if it
were a boat floating (Fig. 11.7). Richard MacCormac has convincingly
argued that this quality of design would have been extremely
unlikely to emerge if the designers had changed between the
outline and detailed design stages as is now common in some
methods of building procurement. Sources of primary generators
In the examples considered so far those constraints have been mainly
radical in function, that is to say, they are considerations of the
primary purpose of the object being designed. The architectural
student groups designing a county administrative building focused
their attention on providing satisfactory working conditions and internal
communications. In general there seem to be three main sources
for primary generators or central design ideas. First, and most obviously
as we have seen, the programme itself in terms of the radical
constraints involved. Second, we might reasonably expect any particularly
important external constraints to impact significantly on the
designer’s thoughts. The design of the Severins Bridge across the Rhine in Cologne, which was illustrated in Chapter 6, is a very
good example of a central design idea emerging from external
constraints. Third, we may expect designers to bring their own continuing
programme or ‘guiding principles’ (see Chapter 10) to bear
on the specific project. This deserves further illustration here.
As we saw in the last chapter many architects have some guiding
principles based around practical constraints. One area particularly
popular during the modern movement was that of structure, with the
notion of ‘structural honesty’ forming an important part of many architects’
guiding principles. Bill Howell (1970) described how his practice
of Howell, Killick, Partridge and Amis developed a philosophy of
building they called ‘vertebrate architecture’ in which ‘the interior
volume is defined and articulated by actual, visible structure’. Howell
showed how this led to a design process in which architect and engineer
worked in close dialogue to develop the anatomy of each building.
At first glance this approach seems rather wilful and, indeed,
Howell (1970) admits that ‘we do it, because we like it’. This suggests
a design process which is guided by a general set of principles about
the role of structure, and in which the primary generator is likely to be
the structural form of the building. The sequence of drawings shown
here, drawn during the design process for Howell’s University Centre
building in Cambridge, rather tend to confirm this (Fig. 11.8). Of
course, such a design process cannot exclude all other considerations,
it is just that they are organised around the primary generative
ideas. Howell describes exactly such a process in his own words:
While thinking about structural economy, the relationship of internal
partitioning to downstanding beams, the relationship of cladding to the
structure, and so on, you are taking decisions which affect the relationship
of the anatomy of the building to its site and to its neighbours.
(Howell 1970)
Of course this strategy is not in some way ‘right’ or ‘wrong’. It simply
worked for this particular designer and created an architecture of a
certain kind which has been much admired (Fig. 11.9). By way of illustrating
this we might consider how Arthur Erikson, who has a very different
set of guiding principles about structure, describes his design
process for his Museum of Anthropology in Vancouver:
As with all my buildings, the structure was not even considered until the
main premises of the design, the shape of the spaces and the form
of the building, had been determined . . . It is only when the idea is
fully rounded and fleshed out, that structure should come into play and
bring its discipline to give shape and substance to the amorphic form.
In that sense it is afterthink. The primary generator and crucial constraints
At this point we should examine the importance of the concept of
constraints. It may not always be obvious that what is important to
a client or a user is not always critical during the design process.
In Agabani’s (1980) study of the way architectural students perceive
design problems one experiment required pairs of students to
design a children’s nursery. After reading the brief and watching a video-recording of the site the students were themselves recorded
as they discussed the problem. The very first recorded comment
from one pair of subjects was to the effect that: ‘the most important
thing is that we are going to have children playing outside’
(Agabani 1980). Now while playing outside is certainly a requirement
for nursery design it hardly seems to be ‘the most important
thing’. However, the same designer continued: ‘so which way
round do you put all the playing areas so that they can wander
around?’ (Agabani 1980). This can now be seen as an assessment
not of what is most important to the client or user but what is
critical to the designer. In this case, orientation of major spaces
towards the protected and sunny side of the site followed by a
consideration of vehicular access was quite fundamental in organising
the overall form. In this sense these constraints are seen by
the designer as crucial in determining form and, therefore, worthy
of becoming primary generators. Making sound judgements on
such things must surely be a matter of experience and perhaps one
of the central skills of good designers.
The life of the primary generator
So far we have seen how both empirical research and the anecdotal
evidence gathered from practising designers suggest that the early
phases of design are often characterised by what we might call analysis through synthesis. The problem is studied not in minute
detail but in a fairly rough way as the designer tries to identify not
the most important (to the client) issues, but the most crucial in
determining form. Once a solution idea can be formulated, however
nebulous it may be, it can be checked against other more
detailed problems. In the experimental studies mentioned earlier
both Eastman’s and Agabani’s results show the combined use of
evolutionary and revolutionary modifications of early solutions. In
the evolutionary phase the designer is really following his or her
nose, gradually modifying the embryonic design as it is tested to
see if it satisfies constraints and is found wanting. Eventually, unless
the design proves totally successful, one of two things happens to
halt this evolutionary phase. Either the general form of the solution
reveals itself incapable of solving enough problems, or so many
modifications have to be made that the idea behind the solution is
lost and abandoned. In either case the designer is likely to choose
the revolutionary step of starting a completely new train of thought.
This is the point where creativity is required rather than ingenuity.
The train of thought is broken and no longer sequential. Some effort
has to be made to look for a new set of problems or a new angle.
In fact the whole primary generator may be scrapped in favour of
a new focus. I have overheard many conversations between design
students discussing their progress, where one will tell the other that
they ‘have just started again’. Such a thing is impossible, the design
process can only begin once, and lessons learned, attitudes developed
and understanding acquired cannot be denied. In this context,
then, ‘starting again’ means looking for a new set of generative
ideas around which to build the next onslaught on the problem.
This brings us as close as we can get, so far, to the centre of design
thinking, for the way in which the designer chooses to shift attention
from one part of the problem to another is central to the design
strategy. In experimental studies we have observed many variations.
Some designers only shift attention when they come to a dead end,
while others seem to deal with several ideas in parallel and we
discuss this further in the next chapter.