Planning Practice & Research

ISSN: 0269-7459 (Print) 1360-0583 (Online) Journal homepage: http://www.tandfonline.com/loi/cppr20

Parametric Design in Urbanism: A Critical

Reflection

Olgu Çalışkan

To cite this article: Olgu Çalışkan (2017): Parametric Design in Urbanism: A Critical Reflection,
Planning Practice & Research, DOI: 10.1080/02697459.2017.1378862

To link to this article: http://dx.doi.org/10.1080/02697459.2017.1378862

Published online: 29 Sep 2017.

Submit your article to this journal

Article views: 6

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=cppr20

Download by: [University of Essex] Date: 02 October 2017, At: 04:04

 Planning Practice & Research, 2017
https://doi.org/10.1080/02697459.2017.1378862

Parametric Design in Urbanism: A Critical Reflection

Olgu Çalışkan 
Faculty of Architecture, Department of City and Regional Planning, Middle East Technical University, Ankara,
Turkey

ABSTRACT KEYWORDS
Parametric modelling run by the explicitly defined algorithms Parametric design; urban
generating synchronically auditable dynamic forms and patterns, design; generative urbanism;
Downloaded by [University of Essex] at 04:04 02 October 2017

has become a prominent method especially in architecture. Though design control and planning
the use of parametric models has got wider in urban design, the
critical reflection on the actual and possible application of the
method in urbanism has fallen limited so far. The paper tends to
relate parametric design with the contemporary understanding of
urbanism with regards to the idea of design control in the the context
of complexity. From this perspective, the actual performance of the
model application in urban context is discussed with the renowned
project of Kartal-Pendik Masterplan (Zaha Hadid Architects) in Istanbul,
Turkey.

Introduction: Emerging Approaches to Urbanism in the Context of

Complexity
Within her influential argumentation on American cities, Jacobs (1961) addressed the local
components of urban form (i.e. streets, building and blocks) as the source of life, and claimed
them as the basic units of diversity within collective urban fabric to be generated in time.
Since Jacobs’ (1961) strong polemic against the modernist fashion of spatial planning and
its disdain on local and spontaneous processes, the renewed recognition of the complex-
ity of cities has led to a particular approach in urbanism. The emerging outlook on cities
would be based on inductive reasoning with a special focus on the generative elements in
design. In parallel to the new perspective, under the influence of spatial complexity based
on the unpredictable adaptation of numerous system components that generate emergent
pattern behaviours on the large scale (Batty, 2007; Johnson, 2001; Portugali, 2011), the
deterministic models to predict the future form of cities have been proved to be failed in
planning, as well. The limited capability to deal with the non-linear dynamics of spatial
growth was also conceived as the crisis of rational comprehensive planning that was used
to rely on Newtonian/linear analyses and probabilistic statistics (Batty & Longley, 1997, p.
74). Recognition of the limits of the conventional approaches to urban complexity ends up
with the new generation of urban studies aiming insight rather than prediction (Ibid.). This

CONTACT  Olgu Çalışkan  olgu@metu.edu.tr, olgucaliskan@gmail.com

© 2017 Informa UK Limited, trading as Taylor & Francis Group
 2   O. ÇALIŞKAN

would inevitably lead to an updated outlook of urbanism based on bottom-up collective

pattern formation and control rather than single-minded pursuit of masterplanning and
total design.1
The emerging approach could be called generative planning and design. Originally coined
by Christopher Alexander in 1968 with a precise definition of ‘generating systems’, whose
parts and rules will (incrementally) create the necessary holistic system properties of their own
accord (p. 66), the concept finds its operational insight with the first design implementation
by John Frazer in the early 1970s. Integrating genetic algorithms into the explorative pro-
cesses of architectural form-finding, Janssen et al. (2002) developed the early morphogenetic
applications of design.2
The urban version of the generative design paradigm was originated again from C.
Alexander and his colleagues’ early studies in the mid-1980s. In his seminal book, A New
Theory of Urban Design, Alexander et al. (1987) addressed a kind of urban growth process,
which is piecemeal, incremental and gradual, creating coherent wholes within urban fabric.
Downloaded by [University of Essex] at 04:04 02 October 2017

Accordingly, Alexander et al. (1987) asserted that the large-scale order would organically
emerge from the co-operation of individual actions (p. 37). For this purpose, they demon-
strated the emergent morphological quality with a group experimentation based on rule-
based generative design procedure. This proposition represented a clear opposition to the
idea of total architecture in urban context.
Similarly, Rowe and Koetter (1975) manifested a similar approach with their polemical
article, ‘Collage City’ which3 claims a pluralist approach for the multiplicity of piecemeal
urban tissues as the assemblage and fusion of different styles and programmes.4 This actually
connotes the notion of design for (urban) complexity.
A few years after Rowe and Koetter’s (1975) urban architectural vision, Kevin Lynch
defined ‘site planning’ with some implicit arguments for generative urbanism. For Lynch
(1981), only the main roads, public facilities and major landscape structure should be
planned. Subdivision of the internal areas (i.e. residential streets, plots) should be made
when the actual demand of building occurs within small sectors (p. 232). From that per-
spective, he speculated ‘the future techniques’ as follows: ‘… the computer may allow us to
design, display and evaluate streams of events rather than a restricted set of frozen stages’
(Ibid., p. 286).
The critical reflections on urban design from the same perspective have apparently pro-
vided a new outlook for spatial planning as well. In this context, Salingaros (2000) defined
‘organized complexity’ as the source of planned morphological coherence. Arguing coupling
as the process of building local connections between urban elements, Salingaros (2000)
basically addressed a procedural (planning) perspective to urban design within complexity.
The broader theoretical framework of this viewpoint is found in Marshall’s (2009) treatise
on cities, design and evolution. By conceptualizing the act of design within the overarching
framework of evolution, Marshall (2009) addressed a new (evolutionary) understanding of
planning and design in urbanism. Accordingly, rather than fully composing the ultimate
collective form in the name of city design, what he suggests is urban ordering via controlling
the individual acts of design within a flexible framework ‘genetically’ coordinated by local
codes5 (pp. 253 – 284).
As the champion of the ‘generative’ approach to urbanism, Mehaffy (2008) describes
generativity of design in terms of the lack of master plans, rigid typologies and the ‘design
 PLANNING PRACTICE & RESEARCH   3

partis’, which all reduce the complexity of pattern formation into simple diagrammatic
schemes used for designating the static plan layouts. Instead, he basically argues the use of
stepwise design processes driven by the sequential collaboration of the participants involved
(Mehaffy, 2008).
The overall trend depicted above does actually go hand-by-hand with the emerging
idea(logy) of planning without plan, which eventually points out a generative approach to
urbanism.6 The developing idea of urban complexity has already shifted the mainstream
interest from plan (blueprint) to rule (design codes), in the literature (Ben-Joseph, 2005;
Lehnerer, 2009; Marshall, 2011; Talen, 2012). This promising branch of discussion suggests
a relevant basis for the possible integration of the rule-based generative methods into urban
planning and design. Nevertheless, despite the clear theoretical perspective built up, the
operational insight of the so-called generative urbanism is yet to be developed. The optimum
application scale (i.e. plot, block and ensemble) of generative urban design is not clarified yet.
Moreover, the emerging perspectives on practice require operational tools and techniques
Downloaded by [University of Essex] at 04:04 02 October 2017

to manage the actual design processes in the context of complexity.

In this framework, computational design techniques suggest a serious methodological
basis for practicing the emerging approach to urbanism which essentially reveals an adaptive
and flexible insight as alternative to master planning (Verebes, 2014). With its increasing
use in current practice and research, parametric models, in particular, represent the most
promising tool for generative urbanism. In this regard, the paper discusses the relevant and
responsive use of parametric modelling in urbanism with regard to the intrinsic relationship
between planning and design. To that end, following a brief description of the technique,
the notion of ‘parametric urban design’ is examined in the light of the leading applications
in research. Then the actual pros and cons of the model efficiency will be questioned with
reference to a real planning application in the context of İstanbul, Turkey.

Parametric Design as a Generative System

In computational term, generative systems imply combinatorial (or configurational) pro-
cedures by which the primitive geometric primitives (i.e. lines, planes and volumes) are
searched through different permutations to achieve large quantity of novel composite forms
and patterns (Frazer, 1995, pp. 14, 15). Though the generative thinking has always been
intrinsic within design thinking in history (Mitchell, 1990), the advanced computational
techniques enabled designers to empower the generative capacity of design act by simulat-
ing endless potential solutions through advanced information processing, and made them
more liberated from the early cognitive constrains in the simulation of complex forms and
patterns (Verebes, 2014, p. 117). In this context, as a kind of generative design method,
parametric design suggests very high variations via fully controlled algorithms. As a tool
that works with the generative algorithms creating hierarchal relations, and involving a
large number of inputs controlled by few processing rules, parametric design models can
be considered within the notion of organized complexity.7. Due its very orientation to the
way the geometric components are linked in certain inter-dependency, in addition to the
key concern of form generation, parametric models are considered in the class of so-called
‘associative design’ as well (Trummer, 2011, pp. 181, 182). That means any change in one
of the parameters of a component within the generative system ends up with a relational
transformation within the other components in parametric models.
 4   O. ÇALIŞKAN

Like the other types of generative systems,8 the parametric models, on the one hand, do
not directly specify a design object as the target-form; but instead, they encode the formation
(the design procedure) that generates geometric variations within the specified constraints
(Dino, 2012, p. 208). Creative evolutionary systems, on the other hand, operate on ‘explor-
atory engines’ of genetic algorithm producing a very large population of solutions which,
in turn, tested and processed against certain fitness function—in an iterative manner akin
to natural evolution—(Bentley & Corne, 2002). Though they fall into the main category
of generative systems like creative evolutionary models, parametric models do not work
through the non-linear recursive systems steered by the automated evaluation and selection
mechanisms. Thus, due to the deterministic behaviour of the parametric systems depending
on the pre-defined relationships of the system components, (unlike agent-based models
in which all the formation units interact in an open and adaptive system), Gerber (2014)
argues parametric models not in the class of generative computation, but as an associative
form finding and optimization technique. Yet, despite being operated a priori (therefore
Downloaded by [University of Essex] at 04:04 02 October 2017

deterministic) definition of algorithmic structures as argued by Gerber (2014, p. 393), the

rule-based procedure of parametric models reveals the quite an unpredictable and emergent
nature of complex pattern formations.
As defined by Sakamoto and Ferre (2008), the parametrics is a technique enabling the
designer to create and transform the design objects holistically via controlling the consistent
(interdependent) relationships between the parts in the whole geometrical system (p. 3).
While a ‘parametric model’ can be defined as ‘a set of (algorithmic) equations that express
a geometric model as explicit functions of a number of parameters’ (Davis, 2013, pp. 15 –
42); ‘parametric design’ is the study of compositional systems by defining the relationships
between the dimensions of the form-components dependent upon various parameters.9
Moretti (1971), p. 207 cited in Davis, 2013, pp. 4 – 42) In addition to the easiness of variation
by the countless design iterations within the models, the geometric accuracy suggested for
controlling the design form is the major potency of parametric design models (Gerber, 2007,
p. 6). With the increasing use of parametric modelling in art and design, the designers are
no longer expected to present a finished form-composition, but a rule-based system that
generates the (spatial or graphical) pattern via certain set of parametric inputs within an
algorithmic framework (Scheurer, 2011, p. 269).
From a computational perspective, the most crucial feature of parametric design is the
associative geometry set by the generative algorithm. That means all the parts and the entire
sub-systems of the form (i.e. network, envelopes and subdivisions) are correlated with each
other, and got depended on common parameters. One single change in a component is
received as a processing input and triggers a chain of reaction in the other associative parts
of the model (Woodbury, 2010, pp. 11, 171). For instance, any controlled deviation in the
value of network density within a parametric urban model may automatically run observed
variations in the plot layout of the fabric synchronically. Controlling the countless elements
in a pattern-like compositional system by algorithm while manipulating its configurational
properties within harmonious morphological variations cannot indeed be easily performed
by the analogue design methods by drawing (Sakamoto & Ferre, 2008, p. 55).
Scripting-based (algorithmic) design techniques radically alter drafting-based (CAD-like)
models in design. Being able to operate with the interactive design components, designers
explore larger solution spaces, which involve higher amounts of alternative forms generated
by different set of parameters. Unlike the creative evolutionary systems,10 parametric design
 PLANNING PRACTICE & RESEARCH   5

models exhibits a hybrid nature combining both top-down structuring (i.e. point grids,
force-lines) and bottom-up articulation of the form-components. Though the parametric
variation affords a very large solution space with numerous deviations retaining the notion
of surprise, the designer has a control over the global behaviour of the system by altering
the morphological parameters (i.e. density, intensity and closeness) and reflecting upon
the visually represented spatial composition and pattern (Figure 1). That means heuristic
still has an active role in parametric design as in the conventional design processes. This
is what essentially differs parametric design thinking from the fully automated generation
and selection processes within the evolutionary design systems (Gerber, 2007, p. 229). Gane
(2004) calls this ‘reasoned ambiguity’ in parametric design the systemized emergence (p. 83).
The basic application of the parametric models is derived from setting an algorithm
relating the basic components of the design composition (i.e. line, point and curve), which
are not graphically represented at the outset. The algorithm does not have to be in the form
of a script coded symbolically, but in the form of a workflow constructed visually (Figure
Downloaded by [University of Essex] at 04:04 02 October 2017

2). In addition to the geometric components, the parametric platform consists of a number
of functions (i.e. multiplication and boolean operations) to be linked by multiple connec-
tions. The complete set of relations between the system components basically involves
input-data and processing operations which generate the visual output(s) (Khabazi, 2010,
p. 20). Through the operation of algorithm, the simple set of numerical data is processed
as the information generating the consecutive geometry (Ibid., p. 20). Therefore, instead of
drawing the line directly, the designer specifies the modifiable attributes of the line (i.e. the
initial point, length and direction) and the model generates its associative geometry by itself.

Figure 1. A sample of parametric production of complex structure based on a sub-object: With a large
number of associative generation and editing operations (i.e. cloning, randomizing, scaling, positioning),
the parametric system enables the designer to create unpredictable complex geometries and control
algorithmically. Source: PlugEllo, 2010, by courtesy of E. Propello, 2013.
 6   O. ÇALIŞKAN

Figure 2. The visual definition of the design algorithm in the parametric platform of Grasshopper. Source:
‘Parametric Urban Design’ Workshop conducted by O. Caliskan, Y. Baver Barut and G. Ongun at METU,
Ankara on 30 May 2014.
Downloaded by [University of Essex] at 04:04 02 October 2017

A spatial design experiment could provide a simple case for an adequate comprehension
of the basic components of parametric modelling.11 In the exemplar project given below,
the designer was asked for the exploration of the possible form-compositions of an urban
extension, parametrically, within the constraints of the given site. The major aim of the
design was to create a flexible urban tissue responsive to the local inputs such as existing
roads and proposed land-use (Stefanescu, 2012).
The core design operation with the parametric model12 starts with defining the major
variables (components), and the desired rule-based connections between them. After inte-
grating the external constraints into the algorithm, the designer adds the generative compo-
nents such as design boundaries and attraction points. Attraction point is a component to
internally vary the whole configuration generated. It is used for differentiating the force of
parametric values (i.e. length, height and density) of the nearby components. After setting
the initial conditions, the designer creates a universal grid to be manipulated by two para-
metric operations, rotation and division (Figure 3(1)). Having been devised in accordance
with the alignment of the existing lines (the main roads), the basic grid is configured to
perform as the low-level order of the future form-composition.
The following steps involve the articulation of the grid. Accordingly, by means of the
attraction points initially located, the grid was deformed. By this way, the grid was accen-
tuated by the focal points defined by the attractors (Figure 3(2)). Finally, the cell/block

Figure 3.  Three-steps generation of grid structure: basic grid (1.), deformed grid (2.) and the final
tessellated grid (By courtesy of D. Stefanescu, 2013).
 PLANNING PRACTICE & RESEARCH   7

structure was generated through tessellation. Based on a special parametric operation called
Voronoi tessellation,13 the optimum cell structure was formed (Figure 3(3)).
With the final operation, designer came up with an internally differentiated lattice-like
grid structure, which can be manipulated by the changing location of the attractors. As fol-
lowed in the given case, the reconciliation of the visual and symbolic representations within
the same domain characterizes the procedural thinking of form composition in parametric
design. During reconfiguration of urban grid, the designer reacted on the emergent form
and tested it with the syntactic analysis to ensure the maximum integration of the overall
pattern through a constant dialogue between the emergent pattern and the algorithm.
After fine-tuning the distribution of the small and large blocks—via manual alternation—
in accordance with the desired centrality pattern, the designer added the final sub-algorithm
for the generation of building blocks as the infill over the grid-cells. The setback distance
and building depth were the two parameters that created the block structure. Designating
the lines of density gradients and associating them with the parametric model of the blocks,
Downloaded by [University of Essex] at 04:04 02 October 2017

the designer accomplished the internal variation within the built fabric and terminated the
design process, accordingly (Figure 4).
The generative features of parametric design, in this sense, are embedded in the closed
system of design algorithm. Though the final configuration might look like a top-down
imposition within the limits of the external constraints, its production implicates a certain
degree of emergence in the context of the individual design process.14 Therefore, instead
of an already foreseen visual image, the algorithmic configuration of the rules and com-
ponents acts like the conceptual basis (pre-structure)15 of the emergent form within the
parametric models.

‘Parametric Urbanism’
Considering both the emerging use of parametric design in urban projects and its intrinsic
capacity for pattern formation, one could question whether parametric design has a poten-
tial to define a new approach in urban planning and design in the name of ‘parametric
urbanism’. In order to call a design paradigm ‘new’ in urbanism, both the characteristic
of its spatial (morphological) and socio-political assumptions should be clarified. In this
sense, the clearest insight on the claim of ‘parametric urbanism’ can be found again in the
writings of P. Schumacher, an architect and architectural theorist.
With reference to his various experimental applications of the method in urban context,
Schumacher (2011b) tends to adapt the parametric approach to urbanism. He suggests
parametric design as a response to the need to establish and maintain a complex order within
the evolving urban field (p. 129). The term, ‘field’, in this context, does actually connotes
to a kind of theoretical continuity between so-called parametric urbanism and landscape
urbanism. Originally coined by S. Allen in 1997 ‘field’ implies a spatial condition which is
open-ended and indeterminate, and therefore, lacking clear compositional fixations as in
the case of the complex self-organized systems in the nature. ‘Field condition’, according
to Allen (1997[2011]), is not figural, but relational. The overall form is characterized by a
‘series of intricate local (relational) rules’ of aggregation rather than geometric systems of
composition (Allen, 2011, p. 128). Therefore, a subtle form of spatial coherence is ensured
through small and incremental transitions and local variations that prevent raptures and
fragmentations occurred. In urban context, the conceptualization indicates a spatial form
 8   O. ÇALIŞKAN
Downloaded by [University of Essex] at 04:04 02 October 2017

Figure 4. The generated urban pattern as the production of parametric design. Note the smooth transitions
in form and density throughout the whole fabric. (By courtesy of D. Stefanescu, 2013).

generated through an extensive network of fields demarcated not by clear lines, but by
overlapping surfaces—of the built tissue and/or terrain—(Ibid., p. 136).
The conception of ‘field’ laid down the basis of the emerging domain called ‘landscape
urbanism’ which, in turn, would condition the urbanistic view of so-called parametricism.
Considering urbanism as the generative processes of landscape surfaces (‘thickened fields’)
interacting with territorial infrastructure and building programmes, landscape urbanism
envisions the contemporary urban form as a kind of relational and rhizomatic system
(Waldheim, 2016a, 2016b, pp. 32 – 47). This basically suggests an opposing interpretation
 PLANNING PRACTICE & RESEARCH   9

of urban form as an architectonic model with the new consideration of ordering systems
generating complex surface conditions for programmatic multiplicities and coherence
(Waldheim, 2016a, p. 25).
The practical application of conceptual model argued by S. Allen (1997[2011]) can be
found in the parametric urban models which suggest a substantial shift in the perception
of urban form. For Schumacher (2011b), the associative geometry generated by parametric
design alters the traditional understanding of ‘space’ through discrete distinction of figure
and ground, with the emerging notion of ‘field’ in the form of overlapping the fuzzy domains
of parametric associations (p. 423). Therefore, the concept inevitably challenges the notion
of ‘district’ as one of the fundamental elements of urban image. While the conventional
sensation of an urban district is characterized by the recognizable borders conditioning
the feeling of ‘inside’ and ‘out’ within the fabric (Lynch, 1960, p. 47), the smooth transition
and the continuous change of/in the morphological fields by parametric (urban) design are
to obscure such demarcation. The settled typologies of urban form defined by the discrete
Downloaded by [University of Essex] at 04:04 02 October 2017

zones and objects are replaced by a new ‘swarm-like’ computer-generated morphology

defined by the relational fields of the associated urban elements—i.e. paths and building
envelopes-.16 (Figure 5) Then the orientation of movement inside the complex fabric would
be informed by the lawful accentuations of the abstract vectors of (morphological) transfor-
mations (Schumacher, 2008). This is basically argued as a positive aesthetic and perceptual
quality of the ‘new’ urban form offered by the so-called parametricism.
To Schumacher (2011) this signifies a noteworthy shift form the traditional idea of the
delineation of urban space—by composing a few objects—to the acts of organization and

Figure 5. The visual form of Boston defined by Lynch (1960)—left—and the conceptual sketch of the
urban form designed by parametric modelling—right. Note that the clear segmentation of the districts via
recognizable paths and nodes is altered by the complex fluidity of fields in which the paths and districts
merge into each other, and the attraction points act as the peaks rather than the nodes. Source: Google
Earth 2013; after Lynch, 1960, 19; and by the courtesy of Zaha Hadid Architects, 2013
 10   O. ÇALIŞKAN

articulation of the complex patterns generated with uncountable building blocks para-
metrically (pp. 421 – 423). Such an approach is basically justified with the contemporary
socio-economic and cultural context. For him, the cities of the post-fordist network society
require flexibility and variation in growth and transformation through continuous com-
munication. The control of the self-regulating urban system can only be managed by the
rule-based computational processes rather than the compositional blueprints (Schumacher,
2010). Behind this statement, there is an ideal model of ‘parametric metropolis’ based on
an analogy of complex spatiality and the variegated order of eco-systems (Schumacher,
2012). However, as one of the prominent figures arguing the productive relation between
complexity and urbanism within the current architectural design circle, Schumacher (2009)
fails to recognize the incremental nature of urban complexity, which is by definition contra-
dictory with the idea of large-scale urban-architecture as addressed by him. In that sense,
the question of how the parametric models can be integrated into the framework of gener-
ative urbanism yet seems to be answered by further discussions. Only after a review on the
Downloaded by [University of Essex] at 04:04 02 October 2017

actual applications of the method given below, would the notion of ‘parametric urbanism’
be claimed to be a relevant statement on urban design.

Current State of Art: Parametric Applications in Urban Design Research

The term ‘parametric urbanism’ is originated from Design Research Laboratory (DRL), a
research programme held in Architectural Association (AA), London. Based on the mul-
ti-scalar systems that cover both building systems and urban fabric (Verebes, 2009), the
methods developed there represent quite an architectural approach to urbanism. Since then,
several design schools have been introducing the new algorithmic methods into urbanism
as an alternative to conventional master planning. In the sampling experiments presented
publicly, it is possible to see an outstanding form-language based on the transformation
of few components within one scheme, which exhibits a noticeable diversity within unity
on the large-scale (Figure 6). This point basically distinguishes parametric urban design
from the analogue urban blueprints mostly based on standard repetitions. In the projects
reviewed, density and urban program are determined as the main design parameters in the
generation of collective form.
What we mainly observe in those examples is a kind of open system that seems con-
sciously avoiding completed compositions with a fixed border condition, as Schumacher
argues (2009, p. 23). As long as the systems are responsive to the neighbouring ‘conventional’
urban typologies via parametric adaptations, such a quality can be taken positive. Such a
typological approach potentially ensures morphological coherence in urban fabric.
In order to provide the strong link between architecture and urbanism, the emphasis
on the generative capacity of the concepts, type and typology, has become a key one in the
context of parametric urban design. Along with Lee and Jacoby’s (2007, 2011) theoretical
argumentation on the use of architectural typology as the guideline of urban pattern for-
mation within a parametric framework, the so-called ‘typological urbanism’ has been more
influential within parametric urban modelling works17 (Figure 7). Yet, the point is that in
most of the models, the coherently configured typological variations are mostly perceived
decomposable through the complex whole-part (or system-to-subsystem) relationships.
Therefore, most of the projects are hardly informative about the possible incremental forma-
tion through the individual applications within clearly identified parts of the whole fabric.
 PLANNING PRACTICE & RESEARCH   11
Downloaded by [University of Essex] at 04:04 02 October 2017

Figure 6.  A parametric scripted model of design through alternative tissues based on the different
scenarios, and the building patterns generated accordingly (Almasri et al., 2009). (By courtesy of P. Sovinc,
2010).

Figure 7. Building prototypes generated within the computational matrices as urban fabric: Controlled
transformation of the parametric variations lets the emergence of hybrid typologies and their smooth
transformation in a coherent manner. The prototypical research by Max von Werz, The Architectural
Association (AA), London (2006/07)—left—and that by OCEAN CN (2009)—right—(By courtesy of C.
Lee, 2017 and T. Verebes, 2017).
 12   O. ÇALIŞKAN

DRL’s systemic search for parametric urbanism is not actually the first attempt towards
a generative approach to urban design. Since the mid-1990s, using evolutionary systems in
design, Watanabe (2002) has developed an inductive method of form generation in parallel
with the evaluation programs testing the fabrics and patterns in terms of their environ-
mental behaviours. Likewise, C. Soddu and E. Colabella, the Italian architects, developed
another generative approach to urbanism on the basis of their early researches about algo-
rithmic art and design since the late-1980s. Introducing a mathematical system to design
the morphogenetic codes, the authors basically suggest redeveloping and/or preserving the
morphological identity of the towns. (Soddu and Colabella, 1995; Soddu, 2002) (Figure 8).
With the widespread use of parametric modelling techniques in design, the notion
of parametric urban design has become an attractive design research topic in urbanism
(Schnabel, 2006). In the early applications of the model, parametric control was limited to
the quantification of the form drawn manually. In these models, the basic motivation was
informing the stakeholders about the changing physical parameters (i.e. built area and FSI)
Downloaded by [University of Essex] at 04:04 02 October 2017

through dynamic 3D-modelling, simultaneously.18 With the following applications, we see

that parametric modelling is employed as a generative tool in addition to its control func-
tion. In this context, with his experimental design project, F. Holik and U. Brederlau (2009)
suggested a model to parametrically organize the plots and the building volumes by the
modifiable parameters (i.e. proportion, density and alignment) in relation to environmental
factors like light and wind. With this model, Brederlau (2011) claimed the possibility of high
variance in structure and dynamic modification of individual components while the total
system-relations are maintained (p. 327) (Figure 9, above). As another developed version,
the ‘parametric urban design tool’ modelled by Beirão et al. (2011) offers an interactive
interface informing the third-parties with the global output, which is updated through
generating and transforming the urban configurations by the parametric algorithm. The
major parametric inputs involved are building density, height, grid typologies and street
hierarchy (Figure 9, below).
As of today, CityEngine software system represents the most developed version of the
parametric urban modelling.19. Being operated on parametric shape grammar,20 the system

Figure 8. The urban fabric generated via the ‘identity generative code’ and some hypothetical sequences
of unpredictably generated traditional building clusters. Adapted from: Soddu, 2002, by courtesy of C.
Soddu, 2013.
 PLANNING PRACTICE & RESEARCH   13
Downloaded by [University of Essex] at 04:04 02 October 2017

Figure 9.  Parametric urban design models designed by F. Holik and U. Brederlau (2009) (dynamic
modification of urban structure—top left—and generative development of urban fabric—top right),
and interactive parametric system designed by Beirao et al. (2011)—network patterns and block envelopes
coded by density—(By courtesy of F. Holik and U. Brederlau, 2013; and J.N. Beirão, 2013).

allows the user both generating the fabric via large amount of data-input—‘rule parame-
ter’ options (i.e. size range and (ir)regularity of lot subdivision and building height ratios)
and controlling the parameter values by directly transforming the drawing.21 Nevertheless,
despite the wide scope of parametric index it has, the model merely operates on composi-
tional geometry and lacks the structural (configurational) parameters in design.
After the period of intensive research of parametric urban design by certain groups,
the master plan as the winning entry for an international competition for Longgang City
Centre, Shenzhen, China in 2008 has already become one of the iconic scheme on the
issue. The project, ‘Deep Ground’, made by Groundlab, a London-based design practice,
actually combined the major research questions of parametric urban design within the
methodical framework of landscape urbanism.22 By the model, the designers aimed to
regenerate 11.8  km2 urban land by creating an ecological corridor keeping the existing
‘urban villages’ within the development area. For this purpose, the design group run a
parametric model that integrated the natural terrain with building compositions (through
the concept of ‘thickened ground’) and created a catalogue of generative building typology
to be in continuation with the existing patterns of the villages (Groundlab, 2008) (Figure
10). Along with a subtle internal variation of building pattern generating various types of
traditional courtyards and open spaces by the parametric model, the alternations of the
‘urban configuration’ were generated for the whole site on different variables such as number
and location of density nodes (Bullivant, 2012, pp. 257, 258). Therefore, despite having a
systemic ‘relational’ structure in itself, the model, consequently, could not alter the settled
perspective of total design by parametric modelling as opposed to the incrementalist and
evolutionist understanding of generative urbanism.
With the large collection of projects supported by comprehensive theoretical discussion
on the so-called ‘computational urbanism’, the recent book of Verebes (2014) provides quite
 14   O. ÇALIŞKAN
Downloaded by [University of Essex] at 04:04 02 October 2017

Figure 10. Masterplan for the redevelopment of Longgang City Centre, Shenzhen, China: The parametric
model developed for the design scheme enabled the controlled variation of building typology in relation
to the density pattern generated over a large urban territory. Source: Groundlab (2008).

an updated view on the use of parametric models in urbanism. It is observed that the pro-
jects presented in the book, basically, fall into the category of prototypical researches mostly
made on the level of extensive urban fabrics. In that current state of art, it is observed that
‘parametric urbanism’ is yet to suggest an operational system to manage the incremental
(trans)formation of the contemporary urban complexity in a bottom-up manner, which was
deliberately addressed by Verebes (2014) in the name of adaptation and emergence. Despite
the declaration of the death of masterplanning (pp. 87 – 117), most of the studies behave as
 PLANNING PRACTICE & RESEARCH   15

masterplans simulating the wholesale formation of urban areas on a dynamic basis, rather
than acting as an organizational tool to control and coordinate the pieces for the multiple
(trans)formation of the complex urban patterns. Though they suggest highly flexible and
diverse morphological structures, the parametric models, apparently, rely on the control
systems which are too holistic to be considered adaptive and emergent in urban context.23
In this regard, one could argue that although the early theoretical perspectives have
paved the way for the possible use of generative systems in urban design, it is rather hard
to observe a full theoretical consistency between the generative approach in urbanism and
the parametric urban design models. Unlike what the theory postulates, most of the current
parametric design models tend to operate on large-scale patterns without providing a gen-
erative recipe to simulate piecemeal and emergent morphological transformations. In that
sense, they eventually fall into the class of deterministic modelling as applied for various
architectural and industrial design applications (Hanafin et al., 2009).24
To elaborate these critical points more explicitly, we will continue by examining one of
Downloaded by [University of Essex] at 04:04 02 October 2017

the actual applications of parametric design in urbanism with a real case in which the model
approach came very close to implementation within planning processes.

Parametric Design in Actual Urban Context: The Case of Kartal-Pendik

Masterplan in İstanbul, Turkey
Considering the body of discussion given so far, we can specify the major challenges of the
method in actual urban context with a set of points and questions as follows:
• Efficiency of the parametric models in design control: The model provides a manage-
able and competent procedure for the micro-domain of design through controlling
the complex geometries. Is this also the case in urban design control process in which
the various externalities (i.e. property relations, regulative constraints) are involved?
• The validity of design flexibility in urban context: Parametric models are techni-
cally welcoming for any change in parameters in design. In this sense, it gets along
with the very nature of form exploration, which mostly relies on the sudden shifts in
design criteria and the principal parameters set in the beginning of the design process
(Stefanescu, 2012). Can the algorithmic design schemes provide enough flexibility for
the post-design processes in planning, as well?
• The mutual relation between the design parameters and planning codes: Are the par-
ametric rules in design responsive to the actual urban codes? Does the model enable
designers to perform with the standard rule-sets in a more creative way to increase the
typological variations and interactions in the built-fabric?
• The mode of representation that the model possibly suggests for an efficient commu-
nication in planning process: Does the algorithmic thinking in design yield a transpar-
ent and comprehensible design language that is accessible for all the participants (i.e.
designers, planners, developers and people) involved?
• The potentiality of generative parametric systems to change the settled perception of
the designers on the role of design in the context of urban complexity: Do the designers
perceive the new tool with the renewed conception and comprehensions of urbanism?
To suggest precise answers to the key questions above, the relation between parametric
design and the actual dynamics of design control in planning will be discussed with reference
 16   O. ÇALIŞKAN

to the Kartal-Pendik Masterplan, one of the biggest parametric urban design projects applied
into an actual planning process so far.
The Kartal-Pendik Masterplan is the winning entry submitted to the international compe-
tition which aimed the redevelopment of a post-industrial site on the east bank of İstanbul,
Turkey in 2006. The project area covers a 550-hectares urban land and it is located between
a major international highway and the coastline. Regarding the size of the site and the
programme of the area (6 million m2 floor space including housing, business and cultural
facilities), one can easily recognize the ambition of the design task to control a large section
of development area via a single scheme.
Based on the experience of the ‘One North Singapore Masterplan’ project dated in 2001
as the very first application of parametric modelling (Bullivant, 2012, pp. 83 – 90), Zaha
Hadid Architects (ZHA), the London-based design office, submitted a masterplan to the
Greater İstanbul Municipality with a very similar technique applied for the competition in
Singapore. The design project basically consisted of a series of schemes based on the idea
Downloaded by [University of Essex] at 04:04 02 October 2017

of ‘soft grid’. By means of a special algorithmic model,25 the project primarily tended to
tie the designed network and the existing urban structure together by stitching the major
incoming lines of connections from the surrounding area (Zaha Hadid Architects, 2006)
(Figure 11(1)). via the parametric design tool, the urban pattern was configured through
optimization of the network structure with the minimized average detour factor and a
dynamic urban grid (Schumacher, 2009, pp. 19, 20).
The street pattern, in this regard, acts as the basis for the articulation of urban fabric in
which the locally bundled lines specified the nodes of programmatic intensity. The nodes
are emphasized by height differences in the form of rising towers locating on the neigh-
bouring islands (Zaha Hadid Architects, 2006b). In this context, the urban grid provides
a flexible and smart framework for the generation of urban composition. The basic design
idea of ensuring the calligraphy of an urban landscape underlies all the compositional tactics
involved (Zaha Hadid Architects, 2006a). In this sense, the urban fabric is dealt as a kind of
network, which transmutes itself by the gradual transformation of the building types (Figure
11(5)). Therefore, the block typologies are not imposed on the scheme, but emerge out of
parametric form-generation (Figure 9(6)). The generation of blocks are again based on
the principle of associative geometry by which the articulation of building-block types are
correlated with the size, dimension and the location of the cells in the fabric (Schumacher,
2009, p. 21). The smooth and subtle transformation of the blocks, which created several
peak points through the fabric and faded into open spaces, is essentially compatible with the
idea of ‘field’ as the major principle of parametricism (Schumacher, 2011b, pp. 421 – 433).
In terms of urban coding, creation of such a continuous calligraphic density composition
compatible with the Turkish urban codes which are merely based on the regulations of
discrete urban blocks is disputable. Since the building codes are specified for each singular
block in Turkey, the translation of the dynamic form of the parametric model into the static
descriptions of the bylaw codes within a master plan turned out to be the major challenge
in the actual process of development planning in this context.
The Kartal-Pendik Masterplan Project represents a special case in terms of the rela-
tionship between (parametric) design and planning. The official development plan of the
area has been configured based on the main principles and the architectonic framework of
Hadid’s design scheme. (see Figure 12) The following process after the design scheme, in
this sense, can be called ‘localizing the design concept’ by converting the design parameters
 PLANNING PRACTICE & RESEARCH   17
Downloaded by [University of Essex] at 04:04 02 October 2017

Figure 11. The major schemes from the Kartal-Pendik Masterplan designed by Zaha Hadid Architects
(2006) (By courtesy of Istanbul Metropolitan Planning and Urban Design Centre, 2011).

into the standard codes of planning (Akın, 2011). Since the different urban blocks were
composed in various density levels by the design model, the proposed form-composition
 18   O. ÇALIŞKAN
Downloaded by [University of Essex] at 04:04 02 October 2017

Figure 12.  The urban composition proposed by the design office (2006)—left—and the official
development plan designating the height and land-use zones in accordance to the design scheme
(2007)—right—(By courtesy of Istanbul Metropolitan Planning and Urban Design Centre, 2011 and Kartal
Urban Development Association-, 2012).

was representing a conflicting behaviour with the Turkish planning system that entails an
equal distribution of densities within a plan area in the name of ‘equality of development
rights’. For one year after the competition held, the Turkish local planners worked hard for
making the design scheme implementable by re-configuring the internal subdivision of
the blocks. While the floor area ratios (FAR) were kept same for each parcel, the density
variation within the design was ensured by the differentiation in height (Sönmez, 2011). To
manage such a complicated and long process in coordination with the parametric design
scheme, the landowners in the area were organized as a society in 2007. Redistribution of
the development rights has been pursued by synchronizing the design team in London
with the planning department of the municipality in İstanbul (Göksu, 2011). During this
process, there have been two major tasks on design coding: First, the dynamic parametric
form was requiring subtle height differentiations in each block even on a single plot, which
was not a common case for the spatial planning practice in Turkey. Second, each operation
made by the cartographer during parcelling out the development rights was obliging to
 PLANNING PRACTICE & RESEARCH   19

drive the parametric model over again in London. Because any particular change in the
size of a block were abolishing the overall system of urban layout which was based on an
associative geometry (Özkan, 2011). Such an extra workload in the design control process
essentially demonstrated the lacking link between the local urban codes and the universal
modelling in parametric design.
In executional terms, the articulated urban fabric that consists of a series of density sur-
faces differing to each other represented one of the major problems of the design scheme.
The design team treated the development site as if a single developer owned the total area.
In fact, there were 290 landowners were located in the district, which made such a holistic
design scheme impractical for the collective production of urban land26 (Ibid.). Those points
inevitably lead the local government to serious hesitations and uncertainties about the way
to utilize the plan in the actual building process. This was due to the lack of the adequate
tools for dynamic morphological control, which would link the highly varied height and
massing parameters of the design to the implementable building codes. To overcome this
Downloaded by [University of Essex] at 04:04 02 October 2017

problem, the same design office was re-commissioned to provide a number of design guide-
lines to coordinate the future individual developments in the area (Ibid.). In this context,
we see a kind of total-design approach that mainly results from the design tool selected for
the production of a ‘big artefact’ called urban fabric.
In order to adapt the parametric design model into the application process, the associa-
tion which is in charge of the development of the site conducted a design research in 2013.
Accordingly, a team of Turkish designers tended to define 3D formation of urban blocks on
the existing plot pattern in accordance with the design codes specified by the masterplan
(Figure 13). As seen in the selected illustrations from the report, the analogue form artic-
ulations are far from the desired compositional coherence suggested by the computational
model that generated the original masterplan as the winning entry of the competition in
2006.27

Figure 13. The proposed form of urban block in accordance with the density and setback codes of the
masterplan—left—: The highlighted blocks indicate the impossibility to create higher blocks suggested by
the plan due to the size of the plots in the given block. The overall form of the area illustrated by the design
research indicates the difficulty to ensure the (parametric) composition that was originally proposed
by Z. Hadid’s design model, within actual planning framework and on the existing ownership pattern.
right—(Source: Kartal Urban Development Association, 2013) (By courtesy of Kartal Urban Development
Association, 2017).
 20   O. ÇALIŞKAN

Conclusion
Considering the complexity of cities and the fundamental tasks of the designer in com-
putational design domain discussed, one could claim that in the case of a common use of
parametric design in urbanism, the conventional role of urban designers would be expected
to shift from creation to control in the formation of planned urban fabrics.28 This does not
mean that the designers would loose their position at all, but they might have to transform
their function in design decision-making. In terms of the parametric methods discussed,
urban designers may be involved in so-called parametric urban design process through the
two major design operations:
(1) Devising the core setting of the generative algorithms by design rules,
(2) Evaluating the emergent design forms in terms of the constraints set by the context
and the preliminary design criteria.
Downloaded by [University of Essex] at 04:04 02 October 2017

In this framework, the spatial design profession is mainly required for its fundamen-
tal knowledge on form (morphology) and its algorithmic description in terms of design
parameters. In this sense, the author does not believe that urban designers should behave
like programmers writing computational scripts, but they are basically expected to ration-
alize design via the explicit design rules and parameters to be integrated in the algorithms
by the programmers accordingly. Yet, there is no doubt that the new profile of the designer
will be required to have the basic operational knowledge of computation for an efficient
involvement in the phase of programming.
Another point making parametric design significant for urbanism is that unlike the static
models disregarding the time dimension, parametric design suggests a dynamic type of
modelling that enables one to simulate and observe the changing form and behaviour of the
design form in time. This is exactly what Lynch (1981) had imagined for physical planning,
as we mentioned above. This point has a remarkable potential to challenge the conventional
urban design procedures based on static blueprints. By means of the real-time simulations
of the generation of urban form with the actual codes, the emergent performance of the
existing rule systems can be assessed efficiently. This would trigger the alternative planning
processes incorporating ordering (coding) and design in one framework. There is no doubt
that such an optimist view about the influential power of the model on urbanism assumes a
well-established and robust methodical connection between the actual processes of urban
land development and planning, and the design model.
Eventually, the existing capacity of computational design, which enables the designers
to control all the parametric dependencies of a complex form signifies a critical question
in the context of urbanism: Should designers create ‘complex geometries’ on urban scale
by themselves via parametric modelling, or should they define the necessary generative
conditions which would ultimately yield the long-term emergent adaptations? Without
intending to suggest a conservative position, we could emphasize the design profession’s
enduring misconception of ‘urban pattern’ as a corporate artefact.29 The holistic applica-
tions of parametric methods in urban design indicate an obvious risk for the convention
of ‘new monumentalism’30 under the long-standing influence of architectural hegemony
on urbanism. Unlike its early version in design history, the new compositional approach
to urbanism would be pursued not by a master builder composing the large portions of
 PLANNING PRACTICE & RESEARCH   21

urban fabric with a pen, but by an architect configuring the fields of a parametricized urban
landscape with an algorithm.
More essentially, through its pre-rationalized structure (design algorithm), parametric
design seems to contradict with the fundamental nature of spatial design thinking conceived
as a non-discursive phenomenon (Hillier, 1996, p. 3), in which the process cannot be entirely
codified by the explicable parameters beforehand. Whether all the design intentions are
practically quantifiable is highly a disputable point, as well.
Use of parametric modelling techniques in urban design has actually provided a very
prolific methodical basis for the new urbanistic approaches such as typological urbanism (Lee
& Jacoby, 2011), associative urbanism (Verebes, 2009), relational urbanism (Llabres & Roco,
2012) and so on. As we know, design in urbanism involves too many parameters, which are
sometimes contradictory and irreducible due to the ‘wicked’ nature of spatial planning31
(Rittel & Webber, 1973). By its definition, algorithms are constrained with certain set of
parameters. That’s why; parametric design models inevitably reduce design into a limited
Downloaded by [University of Essex] at 04:04 02 October 2017

set of working parameters such as density, height, network geometry etc.32 Therefore, in the
current state of art, it may be more plausible to limit the claim of ‘parametric urban design’
with productive exploration of better (i.e. coherent, integrated and diverse) urban patterns
as a design support system, rather than aiming a ‘new urbanism’ out of it.
Besides, the parameters in spatial design are always subject to change and to be replaced
with the new ones during the process in accordance with the co-evolutionary nature of
design thinking (Poon & Maher, 1996). On this basis, the actual difficulty of recodification
of an already established parametric model (Davis, 2013, pp. 27 – 42; Terzidis, 2006, p.
45) represents one of the most obvious limitation of parametric modelling in the complex
and dynamic design control processes like the ones in urbanism. This is exactly what we
observed in the case of Kartal-Pendik Project as an actual application of the parametric
design method in urban context within a specific legal and administrative framework.33
After all, does it mean that the ongoing searches for an integration of parametric models
into urbanism are useless or even risky? The principle answer to such a question would not
be necessarily a negative one by cautiously emphasizing the risk. Recognizing the drawbacks
of the parametric methods in urban design, we can realize the potential use of the models in
urban design-control. In urbanism, the real challenge and potentiality of parametric design
does not lie in the playful exploration of complex form-compositions via the algorithmic
models, but in the intrinsic capacity to control collective urban formation in a relational
structural framework. That means, rather than trying to create a kind of ‘design machine’,
algorithmic design methods should be utilized as design support system within long-term
development control processes in planning. Yet, the examination of the publicly acces-
sible parametric design projects shows that the design rules involved in the process are
not explicitly provided to the third parties. The algorithm itself is not conceived as a final
product to be presented along with the design image. Therefore, a new understanding of
intellectual property and ownership of the algorithm is needed to create a collective way
of thinking in (computational) design34 (Terzidis, 2011, p. 99). In that way, parametric
models can function as the computational basis for the controlled generation of emergent
and collective urban forms instead of designing large portion of cities directly in a limited
time-period. To establish such a perspective, a better integration of the complexity theories
of generative urbanism and parametric design methodology is required.
 22   O. ÇALIŞKAN

Notes
1.  Such a change in motivation is also to be ended up with a radical transformation in the settled
perception of urban planning and design from designing to organization—or in Marshall’s
(2009) term, ordering—.
2.  As a term of bioscience, morphogenesis refers to the developmental process of a organism
based on cellular growth and differentiation. In this sense, it connotes a kind of bottom-up
process in pattern formation as opposed to the top-down processes in design (Alexander,
2002).
3.  Later on, the extended version of article was published as a book by the MIT Press in 1978.
4.  The quotation on the basic properties of their proposal apparently signifies the authors’
generative ideas on city form: ‘it is simultaneously an appeal for order and disorder, for the
simple and complex, for the joint existence of permanent reference and random happening,
of innovation and tradition’ (Rowe & Koetter, 1978, p. 8).
5.  For an example to planning a (hypothetical) settlement only by the local design codes—
without a blueprint—, see: Sorkin, M. (1993) Local Code: The Constitution of a city at 42
(degrees) N latitude, New York: Princeton Architectural Press.
Downloaded by [University of Essex] at 04:04 02 October 2017

6.  For a thorough discussion on the concept, see: Alexander, E. R., Mazza, L., Moroni, S. (2012)
‘Planning Without Plans? Nomocracy or Teleocracy for Social-Spatial Ordering’, Progress in
Planning 77(2), pp. 37 – 87.
7.  Algorithm is basically a step-by-step procedure for reaching a decision on ill-defined problems.
In of design computation, it is the computation procedure defining the type and quality of the
elements and the sequence and timing of the operations involved (Terzidis, 2006, pp. 65, 66).
8.  In addition to the parametric systems, linguistic systems based on the syntactic rules (i.e.
shape grammar) and creative evolutionary systems based on genetic algorithms can be also
considered within the class of generative design systems (Dino, 2012, p. 209).
9.  Mathematically, parameter is an arbitrary constant or a variable in a formulaic expression.
For instance, in y = ax + b, a and b are the parameters implying a constant value determining
a point on the line (James & James, 1968, p. 263). That means though every parameter is
categorically a variable, a variable would not necessarily function as a parameter. In this regard,
what called as the ‘parameter’ as the factor having variational values in a design algorithm
should be mathematically counted as the ‘metric variable’.
10. In evolutionary systems, generation repeats itself on the populations selected and recombined
to achieve optimum solutions in time. A ‘seed-model’ is propagated into the population of
alternative forms by means of a code. Within the emergent population, the most ‘successful’
ones are selected in accordance with a fixed fitness-function. Then the process is iterated till
the ‘optimum’ solution is achieved (Bentley & Corne, 2002; Frazer, 1995, p. 65; Watanabe,
2002, p. 10). In this method, the designer’s main role is to determine the performance rule
(code) system and fitness criteria. Without a need for human input, the rest (generation
process) is run out of the direct control of the designer.
11. The project is designed by Dimitrie Stefanescu with the collaboration with D. Mila in the
master’s degree studio in Ion Mincu, University of Architecture and Urbanism, Bucharest
under the supervision of T. Florescu and S. Guta in 2010.
12. For the project, the designer used a parametric platform, called Grasshopper, which is an
application within Rhino, digital design software.
13. Voronoi is a special technique to create cellular pattern in which the divisions are metrically
optimized in accordance with the objects (called ‘seeds’) randomly distributed on space.
14. As clearly stated by the designer, the actual network configured by the universal grid and the
attraction points, was only apparent (on the visualization screen) at the very late stage of the
process in the project (Stefanescu, 2012).
15. For the original definition of the concept, see: Hillier, B., Leaman, A., 1974 ‘How Design
Possible?’ Journal of Architectural Research 3(1), pp. 4 – 11.
16. With the perception motivation behind its definition, the idea of ‘field’ in Schumacher’s theory
can be associated with the model of ‘force fields’ in the gestalt psychology of Kurt Lewin (cited
 PLANNING PRACTICE & RESEARCH   23

in Rittel, H. (1964 [2010]) ‘Environments’, in (eds. J. P. Protzen, D. J. Harris) The Universe of

Design: Horst Rittel’s Theories of Design and Planning, p. 87).
17. As a recent application of parametric modeling in urban context, Llabres and Rico (2012)
suggest a design model based on the generation of a ‘catalogue-based (building) typologies’
and their parametric distribution on the overall pattern in relation to the specific constraints
of the site.
18. For some examples to the early generation of parametric urban design see: Holistic City Software
(2005), CityCAD, http://www.holisticcity.co.uk/, accessed in March 2012; MODELUR—A
Parametric Urban Design Tool (2009), http://www.modelur.com/, accessed in May 2011;
Skidmore, Owing & Merrill Llp, (n.d.) ‘Parametric Urban Design’, http://www.som.com/
content.cfm/parametric_urban_design, accessed in March 2012.
19. The authors of the software define it as a ‘procedural modelling’. It means ‘3D geometries
and textures are constructed using rules (procedures) instead of labour-intensive manual
modelling’ (Esri, 2011).
20. Parametric shape grammar is a type computational system of generating shapes by the
geometric transformation rules. See, Stiny, G. (1980) ‘Introduction to Shape and Shape
Downloaded by [University of Essex] at 04:04 02 October 2017

Grammars’, Environment and Planning B: Planning and Design 7(3), pp. 343 – 351.
21. Since it is compatible with geospatial data, the model is devised to be utilised in the planning
of actual cities as well.
22. Jose Alfredo Ramírez, co-founder and director of Groundlab is actually the director of the
Landscape Urbanism MA at the Architectural Association, London where most of the above-
mentioned design researches were held.
23. After the personal conversation with S. Marshall in August 2017.
24. Hanafin et al. (2009) differentiates determinist models, which require to start design with a
parametrically manipulateable but already given end-product, from the non-deterministic
ones, which allow designer start with the algorithm without knowing its physical representation
at the outset.
25. The design team used parametric modelling software called CATIA to simulate and optimize
the 3D design composition of the competition area. By the software, the model called the
‘wool-thread’ was utilized as an optimization method for the road network to decrease the
overall network length without increasing the deviation/detour factor. With the model, the
lines of the streets were curved, bent and bundled through various locations.
26. Interestingly, the evaluation committee of the competition states in the jury report that
the project offers a potential to make phasing and land subdivision easy in the course of
implementation. see: Topbas, K., Sorkin, M., Kaptan, H., Tur, E. T., Jumsai, S., Ozkan, S.,
Inceoglu, N., (2006) Kartal Alt Merkez ve Kartal-Pendik Kıyı Kesmi Kentsel Dönüşüm Projesi,
Değerelendirme Kurulu Raporu –Kartal- Pendik Urban Design Project Evaluation Committee
Report-, Istanbul Metropolitan Planning and Urban Design Center (IMP): Istanbul.
27. With some other big urban projects, Kartal Masterplan is currently considered as one of the
failed projects in Istanbul. Turkey. See: Kumbasar, C (2016) ‘Kent Boyle Dönüşüyor’—How
the City is Transformed—, Cumhuriyet, 3 April 2016, p. 8.
28. The author’s argument on urban designers is in parallel to the point made by N. Leach (2014)
about the emerging role of architects in the context of computational design: ‘No longer is
the architect the demiurgic form-maker of the past. Rather the architect has been recast as
the controller of processes, who oversees the “formation” of architecture.’ (p. 154).
29. For a concise discussion on the substantial difference between a corporate object and a collective
entity, see: Marshall (2009, p. 135).
30. With regards to the emerging capabilities of parametric design in urbanism, the term positively
used by P. Schumacher (2011): ‘… a new monumental synthesis is achieved that combines
variegated richness with continuity and coherence: intensive coherence’ (p. 427).
31. For a thorough discussion on the lack of concern of current parametric design methods
about the non operational/wicked problems, see: Konig, R. (2011) ‘Generative Planning
Methods from Structuralist Perspective’ in (eds.) T. Valena, T. Avermadete, G. Vrachiliotis,
 24   O. ÇALIŞKAN

Structuralism Reloaded: Rule-Based Design in Architecture and Urbanism, Stuttgart: Edition

Axel Menges, pp. 275 – 280.
32. As one of the most developed parametric application in urbanism, publicized so far, the
designed model at the ‘Relational Urbanism’ studio at the Berlage Institute in the Netherlands
in 2010, sets orientation, land us, height and traffic as the constraining factors involved in the
parametric framework (Llabres & Rico, 2012).
33. Overreliance on the initially set design parameters despite the emerging variables (i.e. property
pattern) that subsequently involved in the process and radically affects the design morphology.
34. To integrate parametric modelling into the collective and participatory processes of urban
development and design, it is necessary to transform the current limitations—i.e. publicly
inaccessible codes, difficulty to modify the model by the third-parties, inflexibility of the
models to be transformed (Davis, 2013, pp. 27 – 42)—mainly directing from the specifications
of the technique.

Acknowledgement
Downloaded by [University of Essex] at 04:04 02 October 2017

The author would like to thank Miray Özkan, a city planner at Kartal Urban Development Association,
Istanbul for the background information and valuable insight provided for the inquiry on Kartal
master planning process, and Asst. Prof. Dr. İpek Gürsel Dino (METU) for her appreciated feedbacks
at the phase of final revision.

Disclosure statement
No potential conflict of interest was reported by the author.

ORCID
Olgu Çalışkan   http://orcid.org/0000-0003-3104-0286

References
Alexander, C. (2001 [1968]) Systems generating systems, in: A. Menges & S. Ahlquist (Eds)
Computational Design Thinking, pp. 58–67 (London: Wiley).
Alexander, C. (2002) The Nature of Order- Book 2: The Process of Creating Life. (Berkeley: The Center
for Environmental Structure).
Alexander, C., Neis, H., Anninou, A. & King, I. (1987) A New Theory of Urban Design, New York.
(Oxford: Oxford University Press).
Allen, S. (1997 [2011]) From object to field: field conditions in architecture and urbanism, in:
A. Menges & S. Ahlquist (Eds) Computational Design Thinking, pp. 119–143 (Chichester, UK: John
Wiley & Sons). Originally published in Allen, S. (1997) From Object to Field, AD Architecture
after Geometry 127, pp. 24–31 (London: John Wiley & Sons Ltd.).
Almasri, S., Chandra, S., Sovinc, P. (2009) Sahra: denCity Project. Available at http://www.peter-
sovinc.com/002.html (accessed 01 April 2012).
Batty, M. (2007) Cities and Complexity: Understanding Cities with Cellular Automata, Agent-Based
Models and Fractals. (Cambridge, MA: The MIT Press).
Batty, M. & Longley, P. (1997) The fractal city, AD Architectural Design, 129, pp. 46–49.
Beirão, J. N., Nourian, P., Mashhoodi, B. (2011) ‘Parametric Urban Design: Interactive Tools for
Supporting Urban Design Decision Making’, The 29th eCAADe Conference Proceedings,
University of Ljubljana, Slovenia, September 21–24.
Ben-Joseph, E. (2005) The Code of the City: Standards and the Hidden Language of Place Making,
Cambridge. (Massachusetts: The MIT Press).
Bentley, P. J. & Corne, D. W. (2002) Creative Evolutionary Systems. (London: Academic Press).
 PLANNING PRACTICE & RESEARCH   25

Brederlau, U. (2011) Parametric design process in urbanism, in: T. Valena, T. Avermadete & G.
Vrachiliotis, Structuralism Reloaded: Rule-Based Design in Architecture and Urbanism, pp. 343–344
(Stuttgart: Edition Axel Menges).
Bullivant, L. (2012) Masterplanning Futures. (London and New York: Routledge)).
Davis, D. (2013) Modelled on software engineering: flexible parametric models in the practice of
architecture, PhD dissertation, RMIT University, Melbourne, http://www.danieldavis.com/thesis/
(accessed 01 February 2017).
Dino, I. G. (2012) Creative design exploration by parametric generative systems in architecture.
METU JFA, 29(1), pp. 207–224.
Esri. (2011) CityEngine. http://www.esri.com/software/cityengine/index.html (accessed 01 March
2012).
Frazer, J. (1995) An Evolutionary Architecture. (London: Architectural Association).
Gane, V. (2004) Parametric Design—A Paradigm Shift? Unpublished thesis Master of Science in
Architecture Studies, The Massachusetts Institute of Technology.
Gerber, D. J. (2007) Parametric Practices: Models for Design Exploration in Architecture, Unpublished
PhD thesis, Harvard University Graduate School of Design.
Downloaded by [University of Essex] at 04:04 02 October 2017

Gerber, D. J. (2014) Parametric tendencies and design agencies, in: D. J. Gerber & M. Ibanez (Eds)
Paradigms in Computing: Making, Machines, and Models for Design Agency In Architecture, pp.
387–400 (Los Angeles, CA: eVolo Press).
Hanafin, S., Pitts, G. & Datta, S. (2009) Non-deterministic exploration through parametric design.
International Journal of Architectural Computing, 7(4), pp. 605–622.
Hillier, B. (1996) Space is the Machine: A Configurational Theory of Architecture, Cambridge. (New
York, NY: Cambridge University Press).
Holik, F., & Brederlau, U. (2009) Experimental Case Study PORTA BCN. http://www.florianholik.
de/blog/?p=364 (accessed 1 May 2012).
Jacobs, J. (1961 [2000]) The Death and Life of Great American Cities. (London: Pimlico).
James, G. & James, R. C. (1968) Mathematics Dictionary. (Princeton: D. Van Nostrand Company).
Janssen, P., Frazer, J. & Ming-xi, T. (2002) Frazer, evolutionary design systems and generative
processes, Applied Intelligence, 16(2), pp. 119–128.
Johnson, S. (2001) Emergence. (London: Penguin Books).
Kartal Urban Development Association. (2013) ‘Kartal Merkez Uygulama Imar Planı’nın 1/1000 Plan
Raporu, Plan Notları ve İmar Yönetmeliğine Göre Test Edilmesi İşi-’. Testing the Codes of the Kartal
Development Plan in Accordance with the Development Bylaw-. (Turkey: Istanbul).
Khabazi, Z. (2010). Generative Algorithms: Using Grasshopper. http://issuu.com/pabloherrera/docs/
generative_algorithms (accessed 01 December 2011).
Leach, N. (2014) There Is No Such Thing as Digital Design, in: D. J. Gerber & M. Ibanez (Eds)
Paradigms in Computing: Making, Machines, and Models for Design Agency in Architecture, pp.
148–158 (Los Angeles, CA: eVolo Press)
Lee, C. M. & Jacoby, S. (2007) Typological Formations: Renewable Building Types and the City. (London:
AA Publications).
Lee, C. M. & Jacoby, S. (2011) Typological Urbanism and the Idea of The City, Architectural Design
81(1), pp. 14–23 (London: Wiley).
Lehnerer, A. (2009) Grand Urban Rules. (Rotterdam: 010 Publishers).
Llabres, E. & Roco, E. (2012) In progress: relational urban model. Urban Design International, 17(4),
pp. 319–335.
Lynch, K. (1960) The Image of the City. (Cambridge, MA: The MIT Press).
Lynch, K. (1981) Site Planning. (Cambridge, MA: The MIT Press).
Marshall, S. (2009) Cities, Design and Evolution. (London: Routledge).
Marshall, S. (2011) Conclusions, in; S. Marshall (Ed) Urban Coding and Planning, pp. 227–243
(London: Routledge).
Mehaffy, M. (2008) Generative methods in urban design: a progress assessment. Journal of Urbanism,
1(1), pp. 57–75.
Mitchell, W. J. (1990) The Logic of Architecture: Design, Computation and Cognition. (Cambridge,
MA: The MIT Press).
 26   O. ÇALIŞKAN

Poon, J. & Maher, M. L. (1996) Emergent behaviour in co-evolutionary design, in: J. S. Gero & F.
Sudweeks (Eds) Artificial Intelligence in Design ‘96, pp. 703–722 (Dordrecht: Kluwer Academic
Publishers).
Portugali, J. (2011) Complexity, Cognition and the City. (Berlin: Springer).
Rittel, H. (1964 [2010]) Models of and for design, in: J.-P. Protzen, D. J. Harris (Eds.) The Universe of
Design, pp. 107–119 (London: Routledge).
Rittel, H. & Webber, M. (1973) Dilemmas in a general theory of planning. Policy Science, 4, pp.
155–160.
Rowe, C. & Koetter, F. (1975) Collage City. Architectural Review, (Aug), pp. 66–91.
Rowe, C. & Koetter, F. (1978) Collage City. (Cambridge: The MIT Press).
Sakamoto, T. & Ferre, A. (Eds) (2008) From Control to Design: Parametric/Algorithmic Architecture.
(Barcelona: Actar).
Salingaros, N. A. (2000) Complexity and Urban Coherence. Journal of Urban Design, 5, pp. 291–316.
Scheurer, F. (2011). Signal to Noise—What is quality in digital architecture?, in: T. Valena, T.
Avermadete & G. Vrachiliotis (Eds) Structuralism Reloaded: Rule-Based Design in Architecture
and Urbanism, pp. 269–274 (Stuttgart: Edition Axel Menges).
Downloaded by [University of Essex] at 04:04 02 October 2017

Schnabel, M. A. (2006) Parametric Design in Urban Design. http://www.scribd.com/doc/49229213/

parametric-design-in-urban-design (accessed in June 2012).
Schumacher, P. (2008) ‘Parametricism as Style—Parametricist Manifesto’, http://www.
patrikschumacher.com/Texts/Parametricism%20as%20Style.htm (accessed in March 2012).
Schumacher, P. (2009) Parametricism—A New global style for architecture and urban design. AD
Architectural Design, 79(4), pp. 14–23.
Schumacher, P. (2010) The Parametric City, in: Zaha Hadid-Recent Projects. (Tokyo: A.D.A. Edita).
http://www.patrikschumacher.com/Texts/The%20Parametric%20City.html (accessed 01 March
2012).
Schumacher, P. (2011a) Parametricism and the Societal Function of Architecture, Public Lecture, 13
May 2011. (Rotterdam: Berlage Institute).
Schumacher, P. (2011b) The Autopoiesis of Architecture: A New Framework for Architecture. (Chichester:
Wiley).
Schumacher, P. (2012). My Kind of Town. http://www.patrikschumacher.com/Texts/My%20Kind%20
of%20Town_The%20Parametric%20Jungle.html (accessed 01 March 2012).
Soddu, C. (2002) ‘La Citta’ Ideale’: Generative Codes Design Identity, Generative Art Conference,
Milan. http://www.generativedesign.com/ (accessed 1 April 2012).
Soddu, C., Colabella E. (1995) Recreating the City’s Identity with A Morphogenetic Urban Design.
17th International Conference on Making Cities Livable Freiburb-im Breisgau, Germany, http://
www.generativedesign.com/ (accessed 15 April 2012).
Talen, E. (2012) City Rules: How Regulations Affect Urban Form. (Washington: Island Press).
Terzidis, K. (2006) Algorithmic Architecture. (Oxford: Architectural Press).
Terzidis, K. (2011 [2003]) Algorithmic form, in: A. Menges & S. Ahlquist (Ed) Computational Design
Thinking, pp. 94–101 (London: Wiley).
Trummer, P. (2011) Associative design: from type to population, in: A. Menges & S. Ahlquist (Eds)
Computational Design Thinking, pp. 94–101 (London: Wiley).
Verebes, T. (2009) Experiments in associative urbanism. AD Architectural Design, 79(4), pp. 25–49.
Verebes, T. (Ed) (2014) Masterplanning the Adaptive City: Computational Urbanism in the Twenty-
First Century. (London: Routledge).
Waldheim, C. (2016a) Introduction: from figure to field, in: C. Waldheim (Ed) Landscape as Urbanism:
A General Theory, pp. 2–11. (Princeton: Princeton University Press).
Waldheim, C. (2016b) Two: autonomy, indeterminacy, self-organization, in: C. Waldheim (Ed)
Landscape as Urbanism: A General Theory, pp. 32–47 (Princeton: Princeton University Press).
Watanabe, M. S. (2002) Induction Design: A Method for Evolutionary Design. (Basel, Boston:
Birkhauser).
Woodbury, R. (2010) Elements of Parametric Design. (London: Routledge).
Zaha Hadid Architects. (2006) Kartal-Pendik Masterplan Press Description. (İstanbul: İstanbul
Metropolitan Planning and Urban Design Center (IMP)).
 PLANNING PRACTICE & RESEARCH   27

Websites
Esrisri. (2011) CityEngine. http://www.esri.com/software/cityengine/index.html (accessed 1 March
2012).
Groundlab. (2008) Deep Ground, masterplan for Longgang City, Shenzhen, China, http://groundlab.
org (accessed 15 September 2015).
PlugEllo. (2010) C4d Plugin PlugEllo, /?m=m1&l=EN (accessed 15 May 2012).

Interviews
Akın, U. on Kartal-Pendik project, 15.06.2011, İstanbul Metropolitan Planning and Urban Design
Centre, İstanbul, Turkey.
Göksu, F. on Kartal-Pendik project management, 15.06.2011, Kentsel Strateji –Urban Strategy-,
İstanbul, Turkey.
Özkan, M. on Kartal-Pendik project implementation, 15.06.2011, Kartal Kent-Der –Kartal Urban
Development Association-, İstanbul, Turkey.
Sönmez, O. on Kartal-Pendik project development, 15.06.2011, İstanbul Metropolitan Planning and
Urban Design Centre, İstanbul, Turkey.
Downloaded by [University of Essex] at 04:04 02 October 2017

Stefanescu, D. on parametric design systems, 08.06.2012, TU Delft, Delft, the Netherlands.