Contemporary digital design practice is in a state of rapid evolution. While architects have employed
computer-aided drafting (CAD) systems for decades, only recently have two distinct and potent design
sensibilities - parametric and algorithmic design - emerged. Nurtured by early architectural researchers and
programmers operating in practice, these methodologies are now gaining widespread professional and
academic acceptance.
Together these two techniques are opening up a new field of possibilities for architectural practice. Most
significantly, they have been developed and refined primarily in commercial practice and not in academia.
Since the late 90s, these advances have coincided with the emergence of a number of digital research
units within commercial practice, such as the Specialist Modelling Group at Foster and Partners, Gehry
Technologies spun off of Gehry Partners, the Advanced Geometry Unit at Arup, and CODE at Zaha Hadid
Architects. These in-house digital research units have been developed as a means of ensuring that the
complex buildings of today are designed and constructed efficiently, on time and within budget. We can
therefore discern an evolution within the development of digital design tools, from a period when they were
associated with science fiction and virtual reality in the 1990s, through to a period when they began to be
used to understand within the realm of digital tectonics to understand material behaviors in the 2000s,
through to a moment when they have become almost indispensable in the production of complex buildings
in the 2010s.
However, these two design sensibilities – parametric and algorithmic design - are often confused, and
sometimes collapsed into the single term, ‘parametricism’. This article is an attempt to clarify the situation,
and to offer some precise definitions of terms in order to differentiate these two quite distinct digital
techniques. It is also an attempt to evaluate the term ‘parametricism’ itself, and to question whether or not
what we see emerging is a new ‘style’ of architecture [1].
Parametric Design
Parametric is a term used in a variety of disciplines from mathematics through to design. Literally it means
working within parameters of a defined range. Within the specific field of contemporary design, however, it
refers broadly to the utilization of parametric modeling software. In contrast to standard software packages
based on datum geometric objects, parametric software links dimensions and parameters to geometry
thereby allowing for the incremental adjustment of a part which then affects the whole assembly. For
example, as a point within a curve is repositioned the whole curve comes to realign itself. The operations
that it facilitates are adaptation, blending and smoothing. It is therefore useful not only in modeling
individual forms but also in the whole field of associative urban planning.
Parametric software lends itself to curvilinear design, as in the work of Frank Gehry, Zaha Hadid, and
other architects whose work is characterized by the manipulation of form. In itself, however, parametric
software does not open up a new vocabulary of form. Such offices were modeling using analog techniques
long before the introduction of parametric software. However, these techniques are highly efficient
for remodeling forms, and afford greater control in the design process. They also provide more precise
information for digital fabrication processes.
It would be wrong, however, to assume that parametric design is concerned solely with form-making. On
the contrary, parametric techniques afford the architect with new modes of efficiency compared to standard
approaches and new ways of coordinating the construction process (called Building Information Modelling),
as in the case of Digital Project, an architectural version of CATIA customized for the building industry
by Gehry Technologies. The big advantage of such packages is that they allow the construction team to
interface on a single platform, and afford a higher level of control in terms of monitoring the time and cost of
construction.
Algorithmic Design
Algorithmic is a term that refers to the use of procedural techniques in solving design problems. Technically
an algorithm is a simple instruction. It therefore relates as much to standard analog design processes, as
it does to digital design processes. Within the field of digital design, however, it refers specifically to the
use of scripting languages that allow the designer to step beyond the limitations of the user interface, and
to design through the direct manipulation not of form but of code. Typically algorithmic design would be
performed through computer programming languages like RhinoScript, MEL (Maya Embedded Language),
Visual Basic, or 3dMaxScript. In contrast, due to the difficulty of programming, the applications Generative
Components and Grasshopper bypass code with pictographic forms of automation. We might therefore
describe them as forms of graphic scripting. Algorithmic design exploits the capacity of the computer to
operate as a search engine, and perform tasks that would otherwise consume inordinate time. It therefore
lends itself to optimization and other tasks beyond the limitations of standard design constraints [2].
Parametric versus Algorithmic
There is now a widespread practice of conflation the two terms, ‘parametric’ and ‘algorithmic’. This is
partly due to the fact that the two techniques can end up producing similar forms. Algorithmic work
generated using Processing or Rhino script, for example, often has curvilinear forms that are seemingly
similar to work produced using parametric tools. But it is also partly due to the fact that as yet there is little
real understanding of what the terms actually mean – at least on the part of those less familiar with the
world of computation. To some extent the term, ‘parametric’, has become a short hand way of bracketing
much digital design that seems to be curvilinear in its aesthetic expression, thereby providing a convenient
expression for a new style in architecture.
Although there will always be a need for synonyms or accessible terms to describe broad approaches, this
development is a little disturbing, and will lead no doubt to further confusion. Often within architectural
culture terms have been adopted that have little relevance to their use in culture at large. Indeed we can
look through the history of architecture, and find several terms, such as ‘Postmodernism’ or ‘Deconstructivism/
Deconstruction’, adopted within an architectural context to refer to architectural styles in a way that has
made them reduced parodies of their original cultural meanings [3]. Parametrics or parametric design
now seems to be suffering a similar fate: although it is actually a new digital technique that fosters a new
process or methodology of design, it has now been adopted to refer to a new aesthetic expression or style
of architecture.
As computation becomes increasingly prevalent within architectural culture, some effort must therefore be
made to clarify the terms, before the confusion becomes so widespread that the use of the term parametric
to describe a style is sanctioned through sheer popularity. So let us state here quite clearly: algorithmic
techniques are based on the use of code. Parametric techniques are based on the manipulation of form.
They are therefore quite distinct techniques.
At the same time algorithmic techniques are often used in association with parametric techniques. We
might point, for example, to the use of certain algorithmic techniques to generate the initial form that is
subsequently manipulated using parametric techniques. Conversely, algorithmic techniques can be used for
optimization and other operations at the other end of the design process, once the initial generated form has
been modeled through parametric techniques.
There is, however, a deeper concern that destabilizes this neat distinction between the parametric and the
algorithmic, and this is the role of visualization in the process of design. Many architects do not learn to
write code from scratch, but learn by ‘hacking into’ the logic of the code displayed in boxes on the screen, as
they model forms using platforms such as Maya or Rhino. In other words, the code is used merely to provide
the ‘key’ to the manipulation of form, and can be understood primarily through the medium of the visual.
This emphasis on the visual is heightened further with the introduction of recent applications - such as
Grasshopper - that bypass code by using pictographic forms of automation. In the end it is not so clear how
many architects are working within a truly algorithmic framework, as opposed to simply operating within a
visual framework.
Further, it could be argued that parametric and algorithmic operations share certain similarities, in that
both are containers in which values can change based on user input. Just as one can adjust the code
algorithmically to generate different outcomes, so one can adjust the form parametrically to generate
different outcomes. Both operations appear to be based on the adjustment of parameters. Indeed the
introduction of graphic scripting techniques, such as Grasshopper, serves only to blur the distinction
between parametric and algorithmic design. Moreover all parametric design relies necessarily on code. In
other words we find ourselves in a dialectical situation where code and form rely upon one another. There
can be no form without code, and often no code without form. To some extent, then, algorithmic design and
parametric design are merely two sides of the same coin.
Yet there are clear distinctions that can and must be made between the algorithmic and parametric design.
Importantly, it is the duration and permanency of their connections to parameters that distinguish them
as operations. Further, parametric design depends greatly on a manipulation of form that might appear
visually interesting, but is often superficial. The blend-shape modeling operation, for example, can be used
to smooth out form in a seductive visual way – especially at an urban level - but cannot take account of any
useful performance related information embedded in a Building Information Modeling (BIM) model.
Moreover, a distinction should made between graphic scripting techniques, such as Grasshopper, and nonpictorial
forms of scripting. Graphic scripting is relatively limited in its scope. Seldom are there any code trees
to be found in Grasshopper robust enough to compete with the average text based scripting. Indeed graphic
codes have no text editor, and have severe limitations in terms of accuracy, constructability, rationalization
and scalability. Code based scripting enjoys far greater levels of control. Despite their apparent similarities,
then, there are important distinctions to be made not only between algorithmic and parametric operations,
but also between visual scripting and purely code-based algorithmic operations.
Parametricism
The popularity of these new techniques has spawned a new term, ‘Parametricism’. Patrik Schumacher,
a partner in Zaha Hadid Architects and a founding director of the Design Research Laboratory at the
Architectural Association in London, has argued that the term should be adopted for a new ‘style’ of
architecture. In his bold claim he highlights the prevalence of parametric design today, a prevalence that is
underpinned, he maintains, by the popularity of recently developed parametric modeling techniques.
Schumacher begins his article, ‘Parametricism: A New Global Style for Architecture and Urban Design’, by
claiming that parametricism is the new global style for architecture: ‘There is a global convergence in recent
avant-garde architecture that justifies its designation as a new style: parametricism. It is a style rooted in
digital animation techniques, its latest refinements based on advanced parametric design systems and
scripting methods [4].
For Schumacher, moreover, it is a style that has succeeded Modernism as the new global style: ‘Developed
over the past 15 years and now claiming hegemony within avant-garde architecture practice, it succeeds
Modernism as the next long wave of systematic innovation. Parametricism finally brings to an end the
transitional phase of uncertainty engendered by the crisis of Modernism and marked by a series of relatively
short-lived architectural episodes that included Postmodernism, Deconstructivism and Minimalism. So
pervasive is the application of its techniques that parametricism is now evidenced at all scales from
architecture to interior design to large urban design. Indeed, the larger the project, the more pronounced is
parametricism’s superior capacity to articulate programmatic complexity [5].
Interestingly, although Schumacher is careful to distinguish parametric techniques from scripting techniques,
he nonetheless brackets them together as part of a bigger movement that he calls ‘Parametricism’.
In other words, although he does not make the mistake of conflating the two, he invents a new term,
‘Parametricism’, that is defined in such a way that it embraces both techniques. According to Schumacher,
then, Parametricism is not limited to the parametric.
What is perhaps more problematic, however, is the way in which Schumacher defines the parametric itself.
Schumacher speaks about it as though it has become some dominant mode of operation. Were we to chart,
however, the various modes of designing in the contemporary academic environment, we would probably
find the use of genuine parametric softwares, such as Digital Project, to be fairly marginal. The vast majority
might be using explicit modeling techniques, such as Maya or 3D Rhino, and a relative minority would be using
graphic/coded automation of some form, but only a very small fraction – maybe 1% - would be using actual
parametric software. Certainly within Schumacher’s own office – Zaha Hadid Architects – Digital Project is
used extensively, but seldom as a design tool. Indeed its role is largely to control the logistics of construction.
What becomes clear – and this is surely the weakest part of the argument – is that Schumacher seems to
want to appropriate the whole pre-parametric legacy of NURBS and SUB-D geometry rule-sets within his
vocabulary of ‘Parametricism’. By doing this, within the scope of ‘Parametricism’ he effectively lays claim to
all curvilinear forms, whether or not they are generated parametrically. Furthermore, although Schumacher
cites Frei Otto’s form-finding experiments, and even claims that Otto ‘might be considered as the sole true
precursor of Parametricism’, he shows little evidence of actual form-finding techniques in his own work [6]. It
begins to emerge that Schumacher is arguing largely not only for style rather than design methodology, but
also for form rather than form-finding.
This brings us to perhaps the most controversial claim made by Schumacher – that Parametricism is a
new ‘style’ for architecture and urban design. Although this ‘style’ is ‘rooted in’ certain techniques including
parametric techniques and scripting, it cannot be reduced to those techniques. ‘What confronts us,’ notes
Schumacher, ‘is a new style rather than merely a new set of techniques.’ [7] He argues that the techniques
themselves have inspired ‘a new collective movement with radically new ambitions and values’. This in
turn has led to ‘many new, systematically connected design problems that are worked on competitively by
a global network of design researchers’. This leads Schumacher to conclude, ‘Over and above aesthetic
recognisability, it is this pervasive, long-term consistency of shared design ambitions/problems that justifies
the enunciation of a new style in the sense of an epochal phenomenon.’ [8]
Schumacher is correct to use the term ‘style’, even though it has become a somewhat controversial term.
Schumacher would no doubt argue that ‘style’ is being used here in a very different sense to the typical
Postmodern treatment of it as mere representation or appearance. Style here comes to mean ‘effect’ and
should be understood within a deeper historical framework that stretches back to Gottfried Semper and
beyond, where outward appearance is understood as the ‘effect’ of certain underlying processes. Whatever
the techniques used in generating any form, we certainly need to address the question of the resultant
appearance or representation. Moreover, despite the fashionable contemporary interest in process, we
simply cannot escape representation. It is not simply that within a Deleuzian framework ‘process’ and
‘representation’ should be understood as locked within a mechanism of reciprocal presupposition, in that
process feeds into – and deterritorializes – representation, no less than representation feeds into – and
deterritorializes - process. Rather we should understand representation as a direct ‘effect’ of material
concerns that govern the design process [9]. We can never escape style.
Schumacher is also correct no doubt in implying that techniques themselves are always grounded in
a cultural context – with new ‘ambitions and values’. Techniques can facilitate or even ‘invite’ certain
operations, simply because they make it easier to perform them. Just as in the past the adoption of the
parallel motion on drawing boards encouraged the use of parallel straight lines in the design process, so
too today the adoption of parametric software in the computers encourages the use of curvilinear forms.
However, there is nothing in techniques themselves that ensures the selection of such forms. Techniques
themselves are neutral. Rather, as Schumacher observes, it is culture itself that engenders a certain design
aesthetic, and the techniques can merely facilitate the articulation of that aesthetic. Of course, the system
can also work in reverse, in that the sheer demand for a certain aesthetic can encourage the development
of certain software tools – as happened with the customization of CATIA to produce Digital Project. But
techniques themselves are not responsible for generating a design aesthetic.
Finally, Schumacher is also clearly correct in detecting certain common tendencies within design practices
around the world. It is clear that the use of similar tools and techniques is likely to promote an increasingly
homogenized design culture, a tendency that will be exacerbated by the globalization of culture in general,
along with the growing internationalization of architectural practice and education. Just as architectural
practitioners, such as Zaha Hadid or Rem Koolhaas, are designing buildings all over the world, so too
architectural professors, such as Greg Lynn or Roland Snooks, are teaching in multiple schools of architecture
in different continents at the same time. Likewise through the internet and global publications – such as
this catalog - architectural ideas are being disseminated more and more effectively. It is easy to understand
therefore how a new design aesthetic might be spreading across the globe.
Where Schumacher needs to be challenged, however, is in his use of the term, ‘Parametricism’ to describe
this new design aesthetic. Not only is the work described largely not parametric in that it has not been
generated using parametric tools, but also it has its origins in a pre-computational world. Furthermore, it
seems somewhat perverse, given that Schumacher expresses clearly that what he is describing ‘is a new
style rather than merely a new set of techniques’, that he adopts a term, ‘Parametricism’, that is clearly
derived etymologically from a particular technique. We therefore find ourselves in a somewhat confusing
situation where a relatively marginal computational technique has been co-opted to refer to a whole new
style of architecture.
A New Global Style?
Modernist discourse tends to see the world in terms of universals. By contrast postmodern discourse tends
to present the world in terms of multiplicities. As the ‘grand narratives’ of modernity give way to the minor
narratives of Postmodernity, we can detect a shift from universals to multiplicities, from homogenization to
differentiation [10]. We now live in a world of differences, where the belief in universal truths or universals
of any kind has been called into question. As such, the notion that a new universal style has now usurped
Modernism comes across as paradoxically Modernist.
Yet my real concern is that Schumacher’s position is not Modernist, but Postmodernist, as the very
differentiation of Postmodern culture is also highly problematic. As Fredric Jameson has argued – in the
context of the call for Regionalism within architecture – the urge to counter the homogenization of late
capitalism by celebrating ‘difference’ is itself complicit within the logic of late capitalism itself [11]. Difference,
as Jameson notes, becomes another commodity within the marketplace. Rather then overcoming the
homogenizing tendencies of late capitalism, it can therefore be seen to be feeding them.
By extension, it could be argued that other attempts to resist the global – such as anti-globalization – fall
within the logic of globalization, just as attempts to resist brand-name culture, such as the Japanese nobrand-
name store, Muji, fall ultimately within the brand-name logic. Of course, such a subtly dialectical
argument can work the other way. The more universal things become, the more we notice the differences
between them. The universal tower block, for example, is treated very differently across the globe. While
in Eastern Europe it is often derided as an emblem of Soviet Bloc totalitarianism, in China is celebrated as
the emblem of the new China. The global and the local, it would appear, are not distinct categories, but are
locked into a logic of reciprocal presupposition. The global promotes the local, just as the local promotes the
global. It is not surprising, then, that the hybrid term ‘glocalization’ – a mix of the global and the local - has
gained prominence in recent years.
We might therefore argue that Schumacher’s call for a new style to bring to an end the ‘series of relatively
short-lived architectural episodes that included Postmodernism, Deconstructivism and Minimalism’ is
itself at risk of being co-opted by that which it seeks to resist [12]. In other words, just as Deconstructivism
and Minimalism – far from resisting Postmodernism - can be seen to be the products of Postmodernism,
so too Parametricism could be seen to be complicit within the logic of Postmodernism. Is the new style of
‘Parametricism’ not the latest incarnation of the logic of Postmodernism?
Such a radical conclusion may not be so far fetched as it might seem. If loosely we are to define
Postmodernism as an obsession with the scenographic and appearances, we might see in Schumacher’s
repertoire of highly seductive formal terms – ‘continuous variation’, ‘seemless fluidity’, ‘deep relationality’ – a
concern not to overcome Postmodernism but simply to supply it with a new aesthetic language. If, however,
we are to subscribe to the recent thinking of philosophers such as Manuel DeLanda, who has developed
a discourse of New Materialism out of the materialist philosophies of thinkers such as Gilles Deleuze,
we might see that there is a new paradigm on the horizon [13]. This is a paradigm that focuses less on
formal concerns and aesthetics, and more on material behavior and design intelligence. It is a paradigm in
which performance – including both structural and environmental performance - is becoming increasingly
important, and one in which structural engineers, such as Cecil Balmond and Hanif Kara, have become the
new ‘material philosophers’. In the context of these new developments, Schumacher’s emphasis on a new
visual language seems a little dated.
Surely what the world of computation promises is not merely a new style, but a radically new way of
approaching design, where we embed new computational techniques into evolutionary and emergent
systems, and where we breed systems and test them out in real time, so that the diagram becomes the
reality and reality is the diagram. Form should be seen as largely irrelevant within this new horizon. Instead
we should be focusing on more intelligent and logical design processes. Logic should be the new form [14].
Neil Leach
注释
[1] I am deeply indebted to Nick Pisca for his advice on this article. Certain sections were cowritten
with him, and Nick has provided some invaluable feedback on the ideas expressed here.
[2] The term ‘optimization’ should be used with a degree of caution. It is often impossible to
know what the optimum might be, and much depends on the parameters being used to judge
that optimum. As those parameters vary, so the definition of the optimum will also vary. Instead
of thinking in terms of optimized solutions we should perhaps refer to ‘improvements’. A more
appropriate term might therefore be ‘ameliorization’ (from the Latin, ‘melior’, meaning ‘better’),
rather than ‘optimization’ (from the Latin term, ‘optimum’, meaning ‘best’). Here, at any rate, the
term ‘optimization’ is used to refer not to the act of finding the optimum solution, but rather to the
process of searching for it.
[3] In the case of Postmodernism, a word used to describe the aesthetic reflex of Postmodernity,
an era generally categorized by cultural theorists as a moment in which everything is co-opted into
images and commodities, the term was promoted by writers such as Charles Jencks, and used
to describe a new style of architecture. In the case of Deconstruction - a mode of philosophical
enquiry that was premised on the notion of challenging the value-laden hierarchies within Western
metaphysics – the term was adopted by the publishing industry, and somehow became charged -
for a short time at any rate - with describing a new formal language in architecture.
[4] Patrik Schumacher, ‘Parametricism: A New Global Style for Architecture and Urban Design’ in
Neil Leach (ed.), Digital Cities, Architectural Design, Vol. 79, No. 4, July/August 2009, pp. 14-23.
[5] Ibid.
[6] Schumacher, p. 23.
[7] Schumacher, p. 15.
[8] Ibid.
[9] We might argue, for example, that the generic appearance of a bicycle is not a consequence
of some arbitrary aesthetic decision making process, but of the material configuration of the
bicycle itself, as a machine designed to offer an efficient form of leg-powered transportation.
In other words, although some form of stylization is inevitable in the design process, the form
has evolved into a more or less stable state primarily through a consideration of mechanical
efficiency.
[10] On this see Jean-Fran鏾is Lyotard, The Postmodern Condition: A Report on Knowledge,
Minneapolis: University of Minnesota Press, 1984.
[11] Fredric Jameson, ‘The Constraints of Postmodernism’ (extract) in Neil Leach ed., Rethinking
Architecture, London: Routledge, 1997.
[12 ] Schumacher, p. 15.
[13] For a discussion of New Materialism, see Neil Leach, ‘New Materialism’ in Neil
Leach, Xu Wei-Guo (eds.), (Im)material Processes: New Digital Techniques for Architecture,
Volume 2: Students, Beijing: China Architecture and Building Press, 2008.
[14] I am indebted once more to Nick Pisca for this slogan.