Wednesday, 16 April 2008

Open systems: approaching novel parametric domains

Following an extract from an article featured in the new issue of VERB - From Control to Design- Parametric/Algorithmic Design coming out soon. The publication is A synthetic look at the impact of parametric and algorithmic design on architectural practice. Contributors include: Mutsuro Sasaki, Arup’s AGU, Aranda-Lasch, Michael Meredith (mos). Special thanks to Albert Ferre and Michael Kubo.


Some of the most relevant shifts in contemporary architectural discourse and practice are intrinsically connected with evolution in computation techniques and software development.

The novelty in architectural design brought forward by new computational tools is often related to software packages or digital techniques developed in other design fields. Innovations in computational as well as manufacturing processes, in fact, experimented and developed by naval, aero, automotive and products industries have represented seminal undertakings for innovation in the construction industry and, moreover, for experimentation in architectural practice.

The introduction of parametric software packages in the world of architecture and structural engineering, despite being a fairly new paradigm, is already redefining the discipline from within.

Traditional CAD products create lines, arcs, circles and a great variety of geometrical objects; making design changes to a given geometry requires changing all appropriate components in order to make the drawing correct.

A new generation of parametric design systems establishes models defined by a collection of constrained relationship between objects. In other words it allows setting up parametric geometrical arrangements capable to build anticipated variations between objects. To better understand what is a parametric set up let’s imagine, for instance, two circles whom centers are connected by a line. The length of the line is, at a time, the sum of the radiuses of the two circles. In a parametric model, varying one of the radiuses affects the length of the line and the mutual dimension of the two circles along that line.

A parameter thus is a variable to which other variables are related by means of parametric equations: design modification and creation of a family of component parts can be performed efficiently by setting up reconfigurable smart models capturing the underlying logic of the design.

The instrumentation of parametric setups into architectural practice is starting to shift the role of the architect in the design processes: from the design of specific shapes to the determination of those geometrical / algorithmic relationships describing the project and its components. The design shifts from drawing surfaces to setting up rules of interdependency - genotypes - leading to potential differentiation – phenotypes -[i].

The novelty represented by parametric tools in architectural culture hasn’t found architects unprepared to conceptually understand its potential for contemporary practice: the responsiveness by which architects and advanced design firms gathered the resources of associative design has triggered a fast implementation of parametric tools in the software industry as well as an increasing curiosity to apply its potential in contemporary architectural design.

Nevertheless, despite the receptivity of some of the most interesting cutting-edge architectural practices, it is possible to trace certain tendencies concerning different approaches, some limitations and novel developing scenarios.

Part to Whole relationship

Architecture is ultimately characterized by the need for a coherent design logic between different elements forming a whole constituted by an interiority and an exteriority; quality and meaning are achieved through the rigorous determination of all those elements contributing to the interfacing between the different components and the building in its entirety: facades, detailing, proportions, symmetry, modularity just to name a few. Thus, parametric design is important for that: the possibility to establish intricate system of relations between different objects and their properties fusing the hierarchy between parts and whole.

“So far experimental architects have just jumped from top-down determination of parts to bottom-up determination of wholes.”
Greg Ly

What Lynn points out is the delay by which architects have employed parametric design in the development of bottom-up approaches where the determination of components has been prioritized over the design of the whole. Lynn is probably referring to the great production of component-based design or the –often acritical- recent proliferation of pattern-architecture.

This approach, though, has just represented an extreme case. Nevertheless, reshaping the traditional dichotomy between the building and its parts, new digital parametric tools still leave behind some unexpressed potential for contemporary architecture, particularly in relation to the possibility to define highly modulated wholes together with the determination of differentiated non-standard components.

Problem-solving approach

In contemporary construction industry, instead, parametric softwares are often employed in design processes of rationalization and post-rationalization where, given a certain project, the answer to specific problems is required to actualize the desired shape [problem-solving approach]. In this case the potential of computational tools is utilized for its higher degree of precision and speed to deliver tailored ad hoc solutions: the parametric modeling is driven by the need to engineer rational solutions in order to fulfill structural, geometrical or fabrication requirements. In this case, in fact, the potentials for a generative approach are set apart in favour of more pragmatic strategies.

Traditional program packages can initially develop quick and precise 3d models; at the same time any occurring change would imply rebuilding the model over and over again until a fixed determination of the design and all its aspects. This is where parametric models come into place.

The paradigmatic innovation of parametric design originates from its modus operandi: the intrinsic resilience to free-form sketching exercise of Digital Project, in fact, requires a sharper understanding of complex geometry and induces the designer to think through the system logic before even starting to draw a line [iii]. In this sense it is projecting desirable perspectives where architectural design is generated from a set of rules and the interdependent relationships between parts governing the manifold aspects of the design. The advent of parameterization increases the complexity of the design task in relation to the necessity to build up not only the model to be designed but also the conceptual structure that guides the parametric variations.[iv]

Parametric design, on the contrary, despite the more complex and, to a degree, time consuming developmental logic, offers greater advantages: from accommodating unpredictable changes happening during the design process to extracting precise data for structural analysis or fabricators.

From a design point of view it is possible to imagine the advent of design methods based on codified geometrical operations proliferating and interacting to achieve a higher level of complex order: the development of a specific design vocabulary based on parametrically codified instances prefigures a fully integrated design approach where complexity and differentiation emerge from the set-up of coherent and controlled operations.

In this sort of scenario the role of the architect and that of the engineer is contiguous and inform each other in a truly cooperative and generative holistic design process.

Open System: branching structures

A second approach towards parametric design is represented by the attempt to build up a deeper understanding of structural systems as multi-performative design set-ups.

Moving away from the homogeneous standardization of the Modern paradigm, this research, through the generative use of parametric tools, is seeking to investigate open systems as multi-performing, differentiated organizational systems.

In line with the experimentation on branching structures developed by Frei Otto, the research unfolds through a series of exercises aiming to open up a generative approach to parametric design: specificity is achieved through iterative differentiation, adaptation through redundancy, robustness through structural-geometric interdependency.

Understanding architectural design as a process of formation leads to the exploration of a pre-material state of a given systems: namely, the state prior to the crystallization into a specific design form is explored. In this way, open systems act as virtual machines prior to the actualization into a given design scenario.

Methodologically, different paths are followed in an attempt to open up potentials for inclusive performance:

Organizational logic> branching is explored as an organizational system. Different network topologies are analyzed and compared.

Geometrical logic> the geometrical logic of branching is created and developed through parametric tools: Digital project is employed to generate the geometrical structure; in addition differentiation is achieved by the instrumentalization of the defining principle: angle between branches, number of branches, length, displacement of the nodes in space…

An intricate matrix is then emerging from the proliferation of differentiated geometrical operation.

Structural logic> the structure and the stability of the various configurations is analyzed through finite element analysis software [FEA]. Thus running structural analysis necessitates specifying a range of parameters to set up likely structural scenarios.

Running structural tests on differentiated geometrical arrangements is possible to detect certain general behavioral patterns happening during the process of extracting precise data for structural analysis or fabrication.

The possibility to establish interdependent relationships between different system logics contributes to the redefinition of common fitness criteria: each system logic, instead of responding to a specific optimized scenario, informs each other towards a multi parametric performing whole. Geometrical arrangements, spatial affects, structural performance and organizational logic contribute to the formation of the system and its performance-based logic.

The development of such a research shifts the architectural paradigm from a problem-solving to a problem-caring approach where integral design logics contribute to the coherent employment of novel design method.

Marco Vanucci, London, May 2007

[i]Patrick Schumacher – Interview AJ 21.12.06

[ii]Greg Lynn – AD Programming cultures, Wiley Academy, 2006

[iii]Patrick Schumacher – Interview AJ 21.12.06

[iv]Robert Aish, Robert Woodbury – Multi-level interaction in parametric design, Smart Geometry Group.