In 1998, the International Foundation for the Canadian Centre for Architecture (IFCCA) held an ideas competition called the Competition for the Design of Cities. The challenge of the competition was to create a speculative design for a sixteen block area of Manhattan (Figure 1). The IFCCA invited a coterie of internationally recognized architects to participate. Among that group was the team of Morphosis Architects and George Yu.

Morphosis Architects is an internationally recognized architectural design firm. Its principle, Thom Mayne, is a winner of the Pritzker Prize in architecture, a founding member of the Southern California Institute of Architecture and a studio instructor at the UCLA School of Architecture. George Yu is the principle of George Yu Architects, formerly Design Office. He is a recipient of the Canadian Rome Prize, a former project architect with Morphosis Architects, and studio instructor at the Southern California Institute of Architecture.

The Morphosis/Yu team immediately recognized the challenge presented by the enormous site. Designing and managing the development of 16 city blocks of dense urban space would require perhaps more than ten years of work.
Within that period, the project would pass through numerous political, social and economic cycles, and the very context upon which the design would be predicated in 1998 would change significantly before its completion. A norm of architectural design is that a project should be integrated in its site. Achieving that degree of integration in a changing context necessitated an approach that would enable both adaptation of the design to change and engagement of the complex and particular conditions of the site. ¹ The history of cities provided inspiration for a novel approach to the challenges of the design process:

The traditional city depended on the stability of its economic and productive structure, a relative uniformity of social composition and the concentrated political power of oligarchies. These conditions were translated into a stable, homogeneous, hierarchic spatial organization.

It is increasingly difficult to experience the city as a linear process or to operate by limiting variation to reinforce a given geometrical structure. It is precisely the instability of the regimes of flexible economic accumulation and the increasing pace of change that bring the variable conditions of new urban structures to the forefront, putting into question traditional city building strategies.

Therefore: a dynamic process with multi-scenario responses (no fixed solutions) to varied forces, the site is to be organized by a set of independent components influenced by formal mutations arising from their interaction, …, the emerging urban construct is no longer linked to a singular city typology, but a composite of historical and emerging city typologies, simultaneously Cairo, Tokyo, the medieval village, New York City, and Los Angeles, and the new next city. ²

IFCCA Design Process

The Morphosis/Yu IFCCA design approach was organized around three conceptual elements: 1) Independent Components, 2) interactions that created formal mutations of these Independent Components, and 3) multi-scenario responses.
Independent Components were geometric objects suggestive of an architectural use or intent. Mayne and Yu characterize these components as morphological characters.
Coupled with each Independent Component was a notion of how the element might be inhabited, how it would transform as a consequence of an interaction with other Components, and the kinds of conditions the element would “seek”.
Independent Components were named so as to capture these traits. For example, “Noodles” were proportionately long, slender elements that would bend and move between other elements as required to accommodate particular conditions.
“Linkers” were long, thin and relatively small elements that would criss cross between other elements, creating interconnections. Other Components included Warp, Snake, Conquistador, Missle, Bar, Pug, Floater, Bit, Passage, Pod, Hold, and Display (Table ?).

Independent Components

Designs were created by bringing Independent Components together in a site model, having them interact with one another, and then organizing their interactions into or around scenarios. Following rules specified in advance, Designers would place Components in the site model then adjust and integrate them into their local conditions. Figure ? illustrates a group of Noodles interacting with one another and elements of the site model. Linkers are deployed to cross connect Noodles where opportunities exist. Independent Components change shape and configuration – what the architects termed formal mutations – as a result of these interactions. The aggregate configuration of these components was the design. Designers could change the mixture of Components to achieve different design objectives.

Scenario

Scenarios guided the interactions of Components, and the development of designs. The concept of scenario as employed in the IFCCA project is related to the concept of version. However, where a version simply makes the distinction that some artifact is different from and related to another artifact, scenarios capture the contingent relation between the context of design and the artifacts that arise in response. In so far as they are interpretations of a context as it exists, or speculations about how a context may develop in the future, scenarios can be said to bind the subjective state of the Designer with the resultant design. Scenarios organize design activity into focused explorations, often centered around design narratives and key factors. Narratives are heuristic devices employed by Designers to guide the development of design activity over time. For example, a Designer may devise a scenario around the notion of “growth” of an element over time. The design that follows would then employ this notion to guide design work, and emphasize those characteristics that make growth behavior apparent in the design. Factors are properties that give a scenario identity.

%\includegraphics[width=1.9in,height=2.17in]{files/introduction/ifcca-designprocess-1.png} %\includegraphics[width=1.9in,height=2.17in]{files/introduction/ifcca-designprocess-2.png} %\includegraphics[width=1.9in,height=2.17in]{files/introduction/ifcca-designprocess-3.png} %\caption{A schematic overview of the IFCCA design approach. a) A collection of Independent %Components are defined. b) Independent Components are brought together in the context, and %interact with one another. c) Interactions are organized by, or organized into design %scenarios. Scenarios are presented as contingent designs.}

George Yu [YU2006] explains that the starting point for the IFCCA design was the need for a public park in Manhattan. The design team decided early on to situate the park on an elevated plinth. Subsequently, each design scenario took the plinth as its starting point. Three major design scenarios, identified as scenarios 33, 15 and 28, were developed. Each scenario explored design possibilities based on different levels of investment in building cost, floor area ratio, total daytime population, and total floor area for the development. Scenario 28 (Figure \ref{img:scenario28}) had the highest total floor area, followed by 15 (Figure \ref{img:scenario15}) and then 33 (Figure \ref{img:scenario33}). In scenarios 28 and 15, a large cluster of Noodles occupy the south edge of the site adjacent to the Chelsea district. The existing Madison Square Gardens building is replaced with a high density tower Component called a Conquistador, and then a new sports facility is designed in a site further west. The waterfront in both of these scenarios is occupied by recreational buildings and artificial beaches. Scenario 33, increases the open park area over scenarios 15 and 28, and the high density cluster of Noodle components is moved farther East.

Yu recalls that one of the original motivations for taking on the project was to experiment with new parametric modeling software. At the time, Alias/Wavefront had just released the first version of Maya, which offered a scriptable modeling environment that could potentially be used to prototype a parametric approach. However, as Yu explains, learning to use Maya while executing the project proved to be too difficult. The early conceptual work of the project was initiated in physical models. Following initial physical studies, the remaining design work was produced using form-Z. Given that form-Z had no scripting support, the team could only suggest through imagery their design intent. As such, they produced a series of Flash animations that were used to communicate the idea of parametric exploration by showing the transformation of the designs based on changes to the four primary factors.

![files/introduction/MAYNE-IFCCA-SC-33.png] \caption{Scenario 33. Low density scenario with large open park area, waterfront Conquistador. \copyright~1998, Morphosis Architects/George Yu, by permission.} \label{img:scenario33}

![files/introduction/MAYNE-IFCCA-SC-15.png] \caption{Scenario 15. Medium density scenario with relocated Madison Square Gardens, waterfront recreational facilities. \copyright~1998, Morphosis Architects/George Yu, by permission.} \label{img:scenario15}

![files/introduction/MAYNE-IFCCA-SC-28.png] \caption{Scenario 28. High density scenario with relocated Madison Square Gardens, waterfront recreational facilities and Conquistador. \copyright~1998, Morphosis Architects/George Yu, by permission.} \label{img:scenario28}

Strategies for Adapting to Change

The effect of the Morphosis/Yu approach was to render the design as a system of interacting parts. Independent Components were morphological types defined through geometric properties and relations. Components then interacted with other Components in the site, leading to formal mutations and an increasing specificity of the Components to their local conditions. Interactions between Components were guided and organized into scenarios that could be presented as contingent designs. This approach supported change and responsiveness to the conditions of the site in a number of ways.

Independent Components were systemic objects with geometric organization and identity. Noodles for example, were proportionately long and slender. When a Designer modified an instance of a Noodle, they would stretch and bend it in the CAD model, understanding that it should remain identifiable as a long and slender object. Second, rules or behaviors associated with the Component guided and localized its interactions with other Components. For example, given that Components were intended as building-like objects, in most circumstances they would need to maintain minimum clearances from one another to ensure proper access to light and air. As such, placing a Component next to another Component would effectively trigger the setback rule, and prompt the Designer to modify the Component and adjacent Components to resolve the condition. As Independent Components interacted with one another, their forms became more specific to their local circumstances. The local focus of interactions meant that Components could be added and removed from the model without necessitating redesigns of the whole. Third, notions of agency coupled with the Components guided their deployment and interactions in the model. When situated in a particular condition, a Component might seek to improve its state, or respond to changes in the context in some particular fashion. The role of the Designer was to execute those responses in light of the identity of the Component.

Scenarios complemented the exploratory facilities provided by Components by constraining and focusing design activity, providing guiding narratives, breaking the design process into chunks, and enabling reflection. Yu explains that there were four principal factors underpinning the three design scenarios — levels of investment in building cost, floor area ratio, total daytime population, and total floor area for the whole development. The multi-scenario approach organized the exploration of variations on these four specific factors. Particular themes emerged in these scenario explorations: low, medium and high density variations, alternative locations of Madison Square Gardens, waterfront recreational facilities, and so on. We might characterize this emergence of themes as an effective clustering of design variations or factor values. Scenario designs and their corresponding design themes evolved over time. This evolution could be conceived of in terms of a continuous process or an incremental movement (ie. in steps) between different conceptualizations. Regardless, the scenario provides a discrete unit of conceptualization or design that can be extended and developed further. The multi-scenario approach is also an overt exploratory strategy that affords increased opportunities for comparison between designs, reflection and learning. By representing scenarios in different states and at different times, the Designer is provided opportunities to learn by considering the nature of the differences presented, and making connections between Designer actions and results.

The design approach distinguished different strata or areas of concern within the overall process and within the design itself. The development of Independent Components, for example, was a different concern and a different process than the management of interactions between Components and the site, or the development and evaluation of scenarios. Such a separation of concerns would not preclude integrated development of these different areas but, afforded the opportunity to reduce the interdependencies between activities such that they could be executed asynchronously. Such a separation of concerns also allowed Designers to focus on more narrow tasks without having to worry about their immediate interrelation with the larger process or design.

Finally, taken as a whole, the approach facilitated change and adaptation through its focus on process rather than on the product of design activity. In providing Independent Components with identity and rules governing their interactions with other Components, Designers would not have to rethink their individual forms when changes occurred. By organizing designs around the basis of interactions, the design could be continually changed and modified without having to think about the design compositionally. By making explicit the process for executing interactions, the design could be realized faster and changes made with less effort.

Analysis

The Morphosis/Yu IFCCA project is multi-faceted, and responds to concerns located both within the competition and within the discipline of architecture itself. In attempting to understand and describe what we find most compelling about the work, we have approached it from various perspectives and sought to situate our discussion and analysis in appropriate theoretical grounds. Our readings have led us to three research domains that we contend are relevant to the work at hand: systems, Designer cognitive behavior, and parametric modeling. We have argued that the IFCCA approach renders the design as a system of interacting parts. The systems sciences provide us with a general model of systems and their properties, and specify heuristic methods that enable their study. The systems approach to organizational management suggests strategies for improving the performance and adaptability of systems.

The cognitive sciences provide empirical evidence of the limitations of human cognitive faculties. Designer cognitive behaviors develop from those limitations and, in response, Designers have developed characteristic strategies for reducing cognitive complexity when dealing with design problems.

The IFCCA approach employs hierarchy and discrete choice throughout the design process, affording means for reducing the cognitive complexity of design tasks.
Finally, parametric CAD systems enable the modeling of both objects and relations; in other words, systems. There is a natural affinity between the IFCCA approach, which we argue renders the design as a system of interacting parts, and parametric tools that facilitate the modeling of systems.

In the following sections, we provide high level overviews of each of these research domains, characterize the properties of the IFCCA approach through the lens of each of these domains, and then draw from those analyses key themes and principles to guide the development of our proposed parametric modeler. Our analysis will also serve to locate opportunities for further research.