The world around us is evolving. We are living inside evolution. As a practicing architect I find nothing more natural than to look around me and implement relevant changes into my own profession. Taking this seriously means implementing new digital technologies in the very fabric of design methods, from the first conceptual thought and from the first accurately described design proposal. Mass production is soon to be overhauled by the principles of customization, in the form of both industrial mass customization and in the form of distributed small scale household fabrication. Customization, which is the modern made to measure, will change architecture from its very foundations. A completely new esthetic will be the natural outcome of the digital parametric design process that connects the file to factory CNC production methods— a new kind of beauty for a new kind of building. Complexity based on simple rules characterizes the dramatic paradigm shift from mass production to customization. The new kind of building is complex yet systemic in its design method. The new kind of building dramatically enhances the potential of todays architectural expression while keeping strict control on its data, including costs. Truly nonstandard architecture is cost-effective and simply complex.


Nonstandard ; Customization ; Complexity ; Parametric ; Programmable ; Swarm

1. As the world turns

The world is changing. So is architecture, the art of building. Since the world is evolving its communication and manufacturing methods drastically and with increasing speed, architecture will never be the same. I present in this writing a theory and practice of architecture which is based on the principles of swarm behavior. It comes down to the provocative assumption that in the end all building components must be designed to be active actors. Based on 20 years of practice of nonstandard architecture I have come to the conclusion that buildings and their constituting components can no longer be seen as passive objects. This assumption revolutionizes the way the design process is organized, the way the manufacturing process is organized, and the way we interact with the built structures. The new kind of building is based on the invasion of digital technologies into the building industry and into the design process, such as parametric design, generative components, file to factory production process of mass customization, and embedded intelligent agents. Step by step we are balancing the familiar top down control with emergent bottom up behavior. Based on simple rules we rethink on the basic building blocks and we build bottom up bidirectional relationships between all constituting building components. I investigate the effects the paradigm shift from mass production to mass customization may have for the designers mind. When the designer is open for this new reality, architecture will never be the same. I will give here one example to visualize the consequences of a truly mass production esthetic. The Cockpit in the Acoustic Barrier project that was finished at the end of 2005 features 40,000 different pieces of steel, and 10,000 different pieces of glass. Not a single building component is the same in this structure. The radicality of this mass customized specimen of nonstandard architecture equals that of the 50+ year old Mies van der Rohes Seagram building, which is the ultimate esthetic expression of mass produced architecture (Fig. 1 ). Mind you, the Seagram building is beautiful, but I would never fancy to strive for such esthetic again, now it is time to find the proper architectural expression for the actuality of industrial mass customization. Within 50 years the paradigm shift toward customization in any form—not necessarily in the form of double curved geometry—will be the dominant language of (inter)national architects. If my assumption proves to be false after all efforts I have done in the last 20 years to develop the practice of industrial customization in the realized works of my architectural office ONL (Oosterhuis_Lénárd) in Rotterdam, and during the last decade the theory of swarm behavior in various educational and research projects with my Hyperbody Research Group at the Faculty of Architecture of the TU Delft, I will be the first to acknowledge it. But if it proves to be right then I will consume the pleasure of having been an early mover to design and construct buildings according to the new rules of industrial customization, and I will feel satisfied having explored the power and beauty of complexity in due time.

Mass production esthetic, Seagram Building, New York, architect Mies van der ...

Fig. 1.

Mass production esthetic, Seagram Building, New York, architect Mies van der Rohe, 1958.

2. Informed point cloud

As the world keeps turning we will need to redefine the foundations of architecture from time to time. Now more than 20 years have passed since the introduction of the PC, since the emergence of the global Internet, and since embedding miniaturized information technology in our consumer products. Today we have become familiar with remote control, wireless Internet, with intelligent agents active on the internet, with intelligent agents embedded in consumer products like printers, cars and computers, but we have not seen much change in the very building blocks of the built environment as of yet. Neither have we seen much change in the way we design and build our environment. We have indeed developed computer programs to simulate otherwise traditional building materials like concrete, steel, and glass composites in a building information model (BIM). Often the BIM is used to improve known designs, largely because most designers do not take advantage of the BIM to develop new design methods, with the aim to realize designs that are not possible with the traditional design techniques. In a BIM the simulated building components are tagged, the tags containing information on their qualitative and quantitative properties. It sure is an opportunity missed that most architects do not use digital design tools in the early design process. Even at respected universities the students are often told not to use the computer to design. It is my explicit opinion that students in architectural faculties should play in the very design process with all kinds of new digital and social media, from day one. It is very compromising for their design skills to postpone designing with new media until they have reached the master program.

While in our practice the parametric design by and large specifies the design of the relational system of the building components, the freehand sketch—whether executed on paper or with the 3D digitizer—specifies the top-down geometrical force imposed on the parametric design system. Both the system representing the bottom-up generation and the ruling curves from the sketches representing the top-down control must be developed in parallel. The top-down force inserts specific data into the parametric bottom-up system.

Sketching in itself is fine, but do express yourself on your touchscreen tablet, use your smart phone to interact and participate—as I conducted some experiments with interactive lectures at Hyperbody, transpose your design concept immediately into a (Grasshopper) script, such that you can play with the parameters and open up your design process to others. The actual emphasis on the “drawing” as advocated at the TU Delft by Michiel Riedijk is conspicuously counterproductive in this respect. In my view the drawing and the section are nothing more than a “flattened” derivate from the 3D model. The drawing and the section should never be the starting point for any spatial design. Building 3D models must belong to the core skills that students in our era are taught. The 3D model contains all information, while the drawings and the section allow for only a poor restricted view of the spatial conditions. Using new media makes the design process transparent, verifiable and participatory, and allows for a stronger individual expression at the same time. The new media are nothing less than another shell around existing media, expanding your world. New media will not replace the old media of language, thinking, conceptualizing, and sketching, but what new media do is to facilitate you to work inside evolution, such that you will participate as an active player in our evolving society. I want to show a possible way forward, forward to the basics of the profession of architecture (Fig. 2 ).

Mass customization esthetic features 1000 unique windows, Al Nasser ...

Fig. 2.

Mass customization esthetic features 1000 unique windows, Al Nasser Headquarters, Abu Dhabi, architect ONL (Oosterhuis_Lénárd), 2012.

To take that step forward I imagine the built structure to be represented by a point cloud of floating reference points that move all the time like the birds in the swarm. The points of the point cloud are continuously informed to behave. The points receive streaming information, process the streaming information, and produce new streaming information, indeed like the birds in the swarm. Complexity based on simple rules. When the information defining its spatial coordinates that are received is not changing, the position of the point in the point cloud remains stable without any change. Now suppose some parametric data are changing, then the point will act accordingly and change its position, or change any of the other properties the point has been tagged with. The crux of the new kind of building is that all reference points will be informed in a streaming fashion both during the design process and during its subsequent life-cycle. Even if we are commissioned to design for a static environment, we must set up the BIM in such a way that all constituting components can potentially receive, process and send streaming information. The BIM will understand its deeper meaning as Building In Motion. Imagine a sound barrier that unfolds only when there is an actual noise source. No noise, no barrier. The noise informs the barrier to unfold and to form a sound insulating shell around the noise source, for example around a train that passes through the city. A wave of the unfolding shell travels along with the speed of the passing train. When there is no train, which accounts for most of the time, there is no need for a barrier. Everyone despises the ugly fences along our highways and along our train tracks. The strong logic of facilitating streaming data to inform built structures makes me confident that this concept of Building In Motion is completely realistic and will become a dominant framework for buildings within 50–100 years. Let us be prepared for this future, let us make designs so as to feel its ultimate logic and seductive beauty. This is the beauty of complexity based on simple rules. And we will perform further research on the possibilities to embed intelligent information processing tags in all building materials for them to be identified and addressed by wireless senders. Literally, think of pieces of steel, concrete, glass, composites with embedded RFID tags to begin with, with microcomputers later, and with a variety of actuators to come. With the Hyperbody group at the TU Delft I have designed and built several prototypes during the last decade showing the enormous potential for a dynamic architecture. With the Barrier In Motion concept we have identified a functional application for the theory of informed point clouds, which is promising to become the basic building block for a streaming connectivity between all constituting building components. Informed building blocks become the actors in an ecology of interacting complex adaptive systems in the Internet of people and things.

3. Forward to basics

The underlying message could very well be: forward to basics. The basics are the dynamic principles of the prototypical building information model (protoBIM), which will be explained later. The implicit assumption is that the basic building blocks of architecture need to be redefined. It is not bricks and mortar, neither is it bits and bytes exclusively. It is rather the merging of bits and atoms which we are concerned about. It is the merging of the old organic real and the new real, the virtual real. One merges into the other, and vice versa. The new building blocks are informed components, hardware augmented with software, mapped on each individual building block. Each individual building block will communicate in a streaming fashion via embedded tags (RFID) with other buildings blocks, anywhere, anytime, anyhow, anyway, thus propagating a radical meaning shift to Eisenmans ANY conferences. The new meaning takes us from a braindead deconstructivism to the vibrant era of synthetic architecture. Synthetic architecture was not accidently chosen as the title of my first solo exhibition in the Aedes Gallery in Berlin in 1990. Synthesizing architecture means redefining the very building blocks and building up a new language from scratch. Synthetic architecture has ever since been subject to a sequence of evolutionary steps: From liquid architecture (Marcos Novak, 1991) via transarchitecture (Marcos Novak, 1995) and Programmable Architecture (Kas Oosterhuis, 1999) to the notion of nonstandard architecture (Frédéric Migayrou/ Zeynep Mennan, 2003). Nowadays it is known practice among advanced students and young digitally educated professionals to use Generative Components (Robert Aish/Bentley Systems), Grasshopper (Rhino plugin), Digital Project (Gehry Technologies), Processing, or similar parametric scripting software to synthesize the new language of architecture. ONL´s contribution in this field has been to actually build on a larger scale of nonstandard benchmark projects as early as 1997 (Waterpavilion), 2002 (WEB of North-Holland) and 2005 (Cockpit in Acoustic Barrier). ONL has effectively built the connection between the bits and the atoms so as to prove that the direction taken as early as the beginning of the 1990s was the right choice. The forward looking approach has led to a new kind of building based on thoroughly redefined genes of architecture—forward to basics (Fig. 3 ).

Parametric CNC produced building blocks, Hyperbody, 2010.

Fig. 3.

Parametric CNC produced building blocks, Hyperbody, 2010.

Forward since we do not want to look back. We do not look in the back mirror to see what is behind us; we simply look around us and appreciate what we see. Now the deep economic depression is the perfect time for innovation in the architecture and construction business. It is the proper time to rethink the basis of our society, thanks to the Internet bubble and the mortgage crisis. It is the proper time to implement streaming nonstandard made to measure strategies in all businesses related to the building industry from designers to manufacturers. Speaking for myself, it is the perfect time to develop the protoBIM innovation based on the principles of swarm behavior in an effort to inspire software developers to support the new kind of dynamic building. Forward to basics does not mean to step back to what we knew already 20 years ago; that would be back to basics. Forward to basics means redefining our core business, redefining architecture, redefining the building industry, redefining the behavior of built structures, and redefining the interface of buildings. Redefining the very essence of our profession.

4. Unique address for each building component

The very essence for the designer software I am interested in is to see all constructs (buildings, installations, environments) as in principle dynamic structures, consisting of a large set of thousands of programmable components. Programmable components are individuals with a unique identity. They have a unique address, in the same fashion as all computers are assigned unique IP (Internet Protocol) addresses. Only because of this unique IP address each individual computer can be connected as an actor and as a receiver to the global Internet. When a building component has an address, it can receive instructions and can accept information either pulled or being pushed from a database. Receiving, processing and sending data means that this building component becomes an actor, such that it can change its configuration. This has been the basis for the pure invention that is the ONL project Trans-Ports imagined in 1999, and in ONLs proposal for the programmable interior of the International Space Station (Fig. 4 ). The invention is to regard buildings as instrumental bodies, which can change their shape and content in real time. Bodies can be addressed, and all constituting components which make up the entire body can be addressed individually. The building components are like the cells in the body, small processors of information, working together while constituting the character of the building body as a whole. To be more specific, a programmable building component could be an actuator in the form of a hydraulic cylinder with embedded sensors, a structural member which has the capacity to adjust its length by becoming longer or shorter, by adjusting its stroke. In the theoretical yet realizable Trans-Ports project it is calculated that only a limited number of approximately 5×6=30 programmable large actuators is needed so as to evoke the behavior of the dynamic body. The skin of the body would have to be flexible, which is realized by introducing a folded skin loosely fixed to the dynamic structure with the capacity to stretch and shrink. Skins with thickness can be executed like overlapping hard scales, thus enabling the dynamic movements. In the example of the Trans-Ports multi-modal pavilion the skin loosely follows the structure. Many other possible concepts with other shapes and different behaviors can be thought of. From the moment one starts to think of a building body as a dynamic construct, a wealth of new possibilities appears at the designers’ horizon, seducing the designers to become pioneers once again. Not superficially modernist, but in-depth modern and above all actual.

Programmable interior for Space Station, architect ONL (Oosterhuis_Lénárd), ...

Fig. 4.

Programmable interior for Space Station, architect ONL (Oosterhuis_Lénárd), 2000.

5. Need for non-linear software

For the design of complex and programmable buildings a basic condition is to work with parametric software. The concept of parametric design is in itself nothing new; it has been in existence for more than 30 years, originated from the shipbuilding industry. Looking more closely into the achievements of the shipbuilding industry, where the design and building task usually is to build large scale one-offs, is useful for understanding the direction where architecture will be heading for in the coming decades. Customization will be the buzz-word, and architects will base their designs on a variety of series of mass customized one-offs rather than relying on the old school serial approach of mass produced components. This can be achieved when we build our 3D models only in a comprehensive parametric way. Parametric design basically means building bidirectional relations, relationships between each individual building component, no exceptions allowed, not “dead” isolated objects. But mind you, the existing parametric software has its pitfalls. Suppose the designer has built a correct parametric model, but based on certain vague assumptions. Then it is very likely that the designer has to revise the model drastically when some of the original assumptions change. And that is what assumptions usually do during the intense evolution of a design concept, which means that all assumptions must be translated into a parametric value. Literally, every seemingly soft design decision must be modeled as a hard parametric fact, verifiable by numbers. There is another pitfall: now suppose the designer switches to another design rule, and suppose the designer changes the rules while playing the design game. That means that the parametric model will need to be restructured from the beginning, which is an even more drastic feedback loop in the evolution of the design. To work with changing rules during the design process we need new species of software, which must be less hierarchical, less linear, more intuitive, and more immediate instead. The relations between the components will need to be more flexible and more like the members of a dynamic swarm indeed. Non-linear parametric software is badly needed for information architects to be able to work more intuitively.

6. Bidirectional relationships

Let me investigate the implications of parametric logic with a simple phrase: I put a cup of coffee on the table. When we try to describe the parametric relationships between the cup and table, between the I and the cup, and between the cup and the coffee, we get very close to the nature of dynamic parametric design and from there we can take the leap toward the essence of behavioral design leading toward a vision of how the new kind of building may be conceived in the early design phase, what it may look like, and –as we will see later–how it may behave. As I pointed out before, we must see all objects, including the I and individual building components, in principle, as actors, as active players in a parametric world. An actor is different from an object since it has an internal drive to act. Now what drives the cup to be a cup and to reside on the table? What drives the I to put the cup on the table? what drives the table to hold the cup? What drives the coffee to stay inside the cup? And when we dive more into the material characteristics: What components constitute the coffee so as to support its labeling as coffee? What happens in the exchange surface between the coffee and the cup? What forces impose the coffee on the cup? And vice versa: what forces impose the cup on the coffee? What forces are from cup to table, from table to cup? Parametric relationships must always be seen as bidirectional. There is always a balance between pushing and pulling, between being pushed and being pulled. Furthermore, what is the relation between the I and the table, which functions as a sort of destiny location for the cup as seen from the viewpoint of the I? For a quick understanding of the subject I need to emphasize the importance of understanding the nature of the interacting components, I, cup, coffee and table. There is a person, a fluid and some objects involved in this interaction scene, components of different kinds, and yet interacting. All interacting components have an impressive history behind them, making them what they are.

Now replace the I by the designer, the cup by a vertical component (the component formerly known as the wall) and the table by a horizontally stretched component (the component formerly known as the floor), and we are talking architecture again. We need to focus on their geometry in the first place, scrutinizing their bilateral relationships on the level of geometry, and on their behavior in the second place, inserting the geometry and all actors in a serious design game, unfolding in real time evolution.

7. Feeling the force

A parametric relationship must be understood in terms of information exchange. The I informs the cup to be placed on the table. The designer informs the bottom surface of component 1 to be connected to the top surface of component 2. To be able to design software for parametric structures it is crucial to make a complete functional description — a script, a scenario if you wish — of all commands which are set into action to relate component 1 to component 2. The two components need to share a point of reference, separately specified for both components. The points of reference are the active members of the point cloud. Once the points have been defined properly, one may connect the two points so as to share the same coordinates in an agreed coordinate system. Once connected the two components must calculate the area of contact they share. If the bottom part of components is flat it will be the full surface area of the standing part which is shared. This area will be used for structural calculations transferring the loads from standing to lying element. It is not my intention to technically describe what algorithms are running in the parametric software to perform these basic calculations. Ultimately, it is my intention to be empathic to the force fields between the components, so as to feel the forces while designing. Feeling the forces in an empathic and sympathetic way is the prerequisite to be able to elevate the basic technique of parametric design toward the level of behavioral design. One needs to internalize the forces. Information exchange from point to point, from surface to surface basically needs to be seen as streaming information, not just as an instance from a stream. Working with streaming information has an emotional effect on the behavioral designer. Streaming information in relation to the time based instances of 3D modeling is like Spaceland in relation to Flatland, as is the 3D model in relation to the flat geometrical instances of the 2D drawing (Fig. 5 ).

Flatland, A Romance of Many Dimenions, author Edwin Abott Abott, 1884.

Fig. 5.

Flatland, A Romance of Many Dimenions, author Edwin Abott Abott, 1884.

Streaming in both directions, both components need to inform each other continuously about their conditions. For example, when the standing component 1 has varying loads due to changing wind conditions, it needs to transfer the real time dynamic data in a streaming fashion to the supporting horizontally stretched component 2. Think of applying this dynamic concept to a 1 mile high building. Such a high building would sweep several meters to left and right and cause nausea for users of the top floors. Now assume that we build a series of actuators in the load bearing steel structure, which actively resist the changing wind forces, thus leveling out the influence of the winds. Then the one mile high structure will stand perfectly upright without any movement in the top. It will stand like a human balancing in the wind, stressing muscles so as to counter the wind. Such a structure would need to send updates in milliseconds so as to keep track of the changes, allowing the actuating components to respond and reconfigure accordingly.

8. From protoBIM to quantumBIM

BIM is commonly known as the building information model. The 3D geometry defines the wireframe, the surfaces, and the volumes. The object is labeled with properties and their performance is described. Virtually everything which has geometry is organized in the BIM. The ideal BIM is a parametric model, meaning that each individual component has a strictly defined relation to its neighboring components and to its object family. Changing one component means changing the local and global relations between the components involved. Adding one component means creating new relationships. As relations are always bidirectional, both of the components are affected by the relationship. Explained in more prosaic language, the wall stands on the floor, while the floor holds the wall. As all relations are subject to constraints, and as we will point out later in many BIM programs unnecessarily discriminative constraints, not all relations are possible. The main cause of this is that the BIM supporting programs are not written by designers but by technicians. They do not know better than to accept worn-out conventions from the traditional design practice. The problem is in the agreed existence of standard digital libraries. Once an object is labeled as a wall it can never become a door. Once you have chosen for the family of floors, their members can never become a wall. Once building components are defined as separate species in a building catalog, they will be allowed to have only a limited number of relationships with other species. Similar to specification into different species in nature, once a donkey, nevermore a horse; they simply can no longer crossbreed. It is obvious when looking at the images of ONL´s nonstandard architectural practice that these traditional categories have become obsolete. A door becomes a specification of the homogeneous structural shell system. The door is not taken from a library but a further local specification of the structural system itself. According to this approach each designer will breed a specific cellular system for a particular project, while the originating cells specify so as to embrace specific tasks, i.e., to be a door with hinges. But there is always a road back, and one can always return to the cellular state where the moving part was not yet specified to move.

In the search for finding the key to solve the above specification dilemma Hyperbody has developed a program based on the dynamic principles of swarm behavior (Fig. 6 ). The hrg (Hyperbody Research Group) software organizes the behavior of points in space, while these points are provided with characteristics like strength, area, volume, color, and shape. Positive strength means attraction; negative strength means repelling the points they are linked to. The swarming points are typically represented by vaguely outlined dots so as to avoid clinging to a specific esthetic preference in the early design phases. Nonstandard information architects are aware of the fact that platonic geometry cannot be the starting point for their designs. They must move deeper into the genes of the design materials. The relations between the points of the informed point clouds in digital space give structure to the early design concept naturally in weightless space so as to be able to introduce the forces of gravity in a later stage, in order to avoid the dominance of the ground level. Hyperbody has teamed up with ONL to develop special software for such early design phases. It is common knowledge that it is the earliest conceptual design phase that is the major driving force for the potential of any project. The very first design decision has far bigger impact than all subsequent design decisions. The software that ONL/Hyperbody is in the process of developing is named protoBIM. ProtoBIM supports the development from a written conceptual statement via a swarming behavioral point cloud toward a BIM that contains all required data for building approval and the tender process. The importance of clearly describing the conceptual statement should not be underestimated. A statement that is described in vague non-verifiable terms is bound to compromise the inherent elegance of the design process. The protoBIM connects all relevant disciplines in this early design phase to each other in the most effective and simple fashion. There will be no more data exchanged than is strictly necessary. The structural engineers do not need a complete 3D model from the conceptual designer; they would rather see a simple wireframe which they can import in their specialist calculation software, most likely applying finite element methods. ProtoBIM is not yet supporting streaming information, as is the main feature of a next level Hyperbody software that I have baptized quantumBIM, which basically is protoBIM with additional features supporting streaming data on all data exchange levels based on the same principles of swarm behavior. ProtoBIM communicates via a dynamic database with other programs, but only in quantumBIM the cells of the database will be continuously updated in a streaming fashion, feeding the actuating building components. QuantumBIM is prepared for the foreseen paradigm shift from static to dynamic modeling, which will be facilitating truly dynamic structures being addressed in real time and proactively acting in real time. ProtoBIM supports truly nonstandard architecture while quantumBIM facilitates truly dynamic structures.

Swarm behavior forms the basis for protoBIM and quantumBIM.

Fig. 6.

Swarm behavior forms the basis for protoBIM and quantumBIM.

9. One building one detail

One building, one detail. I have introduced this challenging phrase in earlier writings (paper for Nonstandard Praxis, MIT conference, 2004). Without any reservation I declared: Mies is too Much! Radicalizing the minimalist tendency of Mies van der Rohe, I observed that Mies still needed many different details to prove his point that less is more. His less is still too much. His less is an imposed less in visual appearance, but still a more in number of details. To perform better one single parametric detail must be mapped on all surfaces, which is subject to a range of parameters rendering the values of the parametric system unique in each local instance, thus creating a visual richness and a variety that is virtually unmatched by any traditional building technique. Such visual richness was naturally apparent in indigenous architecture, all made by hand, based on simple procedures. Now the new parametric and customization techniques allow such visual richness on the grand scale of large buildings, which is complexity based on simple rules. Complexity is the real more, based on the truly less. Please be aware of the double meaning: I do respect Mies van der Rohe to the max, which prohibits me from copying or varying the original—it was deliberate violation indeed when Rem Koolhaas forced the Barcelona Pavilion to bend in the early days of his career. Rather, one should endeavor to radicalize Mies instead; one should take the next step forward, instead of looking backward in such incestuous operations. The parametrization of the leading building detail implies an extreme unification; it requires an uncompromising systemic approach, thus allowing for a rich visual diversity at the same time. Les extremes se touchent. The coherence of parts in a parametric design system does not necessarily lead to a harmonic relationship between the parts as suggested by Palladio over 500 years ago, neither as suggested by Vitrivius 2100 years ago. Coherence of parts in a 3D parametric design system covers a much larger bandwidth of possible variations (Fig. 7 ).

Simply complex, iWEB, Delft, architect ONL (Oosterhuis_Lénárd), 2002 (first life ...

Fig. 7.

Simply complex, iWEB, Delft, architect ONL (Oosterhuis_Lénárd), 2002 (first life as Web of North-Holland), 2007 (second life as the iWEB).

The strategy to induce tension by introducing opposing poles, which will be further dwelled upon in the make of that body section, is applicable on the design attitude toward generative detailing as well. Not only did I introduce opposing poles in master planning (Manhal Oasis), in the body plan of building bodies (Saltwaterpavilion, Space Xperience Center), but in the generic structure of the basic architectural detail as well. The parametric detail is generated just by executing a simple rule while retrieving local data for each individual node. Simplicity is thus intrinsically tied to multiplicity. Its intelligence is embedded in the swarm behavior of the node, the programmable dot of the informed point cloud. I applied the above one building one detail strategy in the design for the Web of North-Holland. The whole construct consists of one single but elaborate detail. All details, including the two giant doors, are members of one big family, as described by one single script (Autolisp routine) mapped on the points of the point cloud as distributed on the doubly curved surface of the emotively styled volume.

The one building one details strategy applies to other scales as well. It applies to product design, to which our architectural approach tends to be very close; it applies to urban design as well, the urban blocks representing the building blocks. Recently we have seen convincing examples of parametric urban design proposals designed by Patrick Schumacher of Zaha Hadids office. As early as 1996 my office ONL developed a fully parametric design strategy for the Reitdiep extension to the City of Groningen, for an urban area hosting 1500 homes. I named the design strategy the Attractor Game. It was in its base an open design system that could be played both intuitively and intentionally, setting the location, the strength and the area of influence of the active urban building components.

10. Just there just then just that

I say no to columns, beams, doors and windows from a standard catalog. Instead of making a tasteful selection from the building catalog and becoming a elitist connaisseur of high culture, I am in favor of designing and building project specific building components, for every new building a new consistent set of interlocking building components. It requires no further explanation that the giant door in the WEB of NH, which is basically a cut-out of the building body, is a door in the WEB of NH only. It cannot be applied in any other design; it belongs there, does not fit anywhere else, it forms the intrinsic part of that design. Just there, just then, just that. It is the logical consequence of mass customization that an end product like a door from the standard catalog will not fit anywhere in the body. In this context I must seriously criticize the buildings of Gehry. From a distance one would be tempted to see them as sculpture buildings, but at closer investigation they are not like that at all. All Gehry´s designs are based on traditional spatial planning, like arranging box-like spaces and wrapping them in the upper floor levels with a decomposed arrangement of loose fragments. Doors, windows, and entrances are as traditional as ever, 100% based on the technology and esthetics of mass production. There is nothing nonstandard about it. Gehry as many of his peers has not been willing to loosen the strings to the traditional building industry; they always rely on stylish catalog products for the majority of their buildings components. They still consider mass production as beautiful. Even when the exteriors of their designs use the metaphor of the nonstandard, their insides are full of column grids, beams, doors, walls and windows, all straight from the catalog. They mistake the complicated for the complex. Decon designs are complicated indeed; they need a stack of different details while nonstandard architecture is complex, based on one or only a few different details, all members of a parametric family. Decon modernist building logic is typically wasting its resources, while nonstandard logic is exploiting resources in a more efficient way. Decon modernist style relies on mass production, nonstandard on industrial customization. The essence of the nonstandard is that each and every building component is precisely defined in the design stage, CNC produced, and hence in principle unique in its shape and dimensions. Each building component possesses a unique number to be addressed by the design and engineering scripts. A building component typically is defined as a 3D parametric component that lives in a spatial relationship toward its neighboring components. Just to remind you, the information that is contained in a 2D drawing can by definition not give you such information, since the drawing does not refer to components at all, but only to their 2D “flatland” shadows of their 3D genetic information (Fig. 8 ).

Complicated not complex, Stata Center, Cambridge (USA), architect Gehry ...

Fig. 8.

Complicated not complex, Stata Center, Cambridge (USA), architect Gehry Partners, 2004.

11. Chicken and egg

What came first, the chicken or the egg? My answer to that is just as simple as it is effective: the chicken and the egg are two instances of the same system, meaning that in each stage of development of the chicken–egg system there were both the chicken and the egg. Naturally, neither the chicken nor the egg were worthy of that name in their early development phase because they were not that much specified when they were busy developing the earliest versions of the adaptive chicken–egg system. Chicken was more something like a worm, and hardly to be distinguished from its eggs. I assume that self-copying and giving birth were equivalent events before the chicken–egg speciation process took off. Similar to the chicken–egg problem there is a causality dilemma between nonstandard designs and computer numerical controlled fabrication the nonstandard design being the chicken, CNC the egg. While the nonstandard design is fully controlled by a parametric logically consistent system describing precise positions, dimensions and geometry of each individual unique component, the execution process—in bio-lingo this may be referred to as the offspring—must follow the same logic. Exact parameters drive the design model. The same exact values as those extracted from the 3D BIM by means of automated procedures (Autolisp routine, scripting) must feed the production process. There may not be a quick and dirty translation, or a remodeling, which always will turn out to be a re-interpretation, and there may absolutely not be any human intervention in the nature of the data, which is bound to be the cause of many possible inconsistencies and inaccuracies. Nothing may be lost in translation. The chicken can only produce and lay her own egg herself, and the egg cannot be produced and assembled by another party applying another systemic logic. In the ONL design and build practice it is observed at times that the occasional mistakes that occurred were always due to erratic human interventions in the file to factory process. Human interventions are bound to blur the consistency; the sloppy accuracy and emotional logic of human measuring or counting simply does not match with the machine logic. Do not worry, I am not trying to exclude people from the process. Humans do play the leading role in establishing the concept, in making intuitive choices from a vast multitude of possibilities, in declaring what is beautiful, basically in every aspect of the design and the building process where communication with other human beings is crucial. But mind you, humans are not good at counting, not good at complex calculations, not good in the consistent application of procedures, and not good at working overnight. People are always tempted to rethink a procedure while executing it, to rethink a process while running it, typically changing the rules while playing. Also their brains are very slow in calculations, so much slower than the personal number crunchers, the PC mates. In order to catch up with the current societal complexity, which is an ever expanding evolutionary process, the information architect had to develop machinic extensions, exo-brains, exo-memories, exo-hands, exo-arms and exo-bodies to design and execute the nonstandard designs. That is why nonstandard design and file to factories production are two sides of the same coin. There would not exist a truly nonstandard design without CNC production, and there cannot exist chicken without eggs, or eggs without chicken.

12. New role of the nonstandard architect

Every traditional intervention in the direct link from nonstandard design to CNC manufacturing would compromise the nature of the nonstandard design. Examples of such compromises are seen in the making of the Water Cube and the Birds Nest for the Olympic Games 2008 in Beijing, and in my own practice I am subject to a similar fate created by the predictable traditional attitude of the project developer of the CET project in Budapest. In all these cases the main contractor has chosen to weld the steel structure, hence compromising the accuracy of the structure, and thus breaking the logical link from the complex geometry to a possibly advantageous and consistent file to factory production of the skin. Once compromised, once the chain is broken, all future steps from there forward can no longer be relinked to the CNC logic of mass customization. The process is killed, the egg is not leading to another life form, the umbilical cord is broken prematurely. Needless to say that each instance where the logical chain is broken is representing a major threat to the practice of nonstandard architecture since the client might see only the blurred outcome and blame the inaccurate compromised details on the nature of the nonstandard design itself. But then again, can the contractors and the project developers be blamed to rely on their traditional experience, which is largely based on traditional bricks and mortar buildings? For them the nonstandard logic may not be logical at all, they are presumably not familiar with the advantages of the file to factory process since they have not mastered this process. It is unknown territory for them. Because of the reality of this situation the nonstandard designer will need to rethink his contractual position as a consultant only and will need to take on financial responsibility concerning the manufacturing process. Since nonstandard architects like myself have full control and full confidence that their data are correct and accurate, they must take on the responsibility for the engineering of the geometry, and naturally must be paid proportionally for this responsibility. Since nonstandard architects are among the few parties to have a full knowledge of how the CNC production procedures have embedded part of the logic of the design itself, they should be remunerated to take the responsibility for managing the direct link from design to engineering as well. The benefit for the building industry will be huge: no more mistakes in the correctness and transfer of the data, no more delays in the exchange and understanding of the concept, remodeling will no longer be necessary, production will be clean and precise, assembly always correct, all steps in the design and building process will be just in time, and just what is needed. No more waste of time and materials, the building site will be clean, while recycling can be developed to cover all used materials. There is one important condition though: all production must be computer numerical controlled, all components must be prefabricated, including all concrete structures and the foundations. Now suppose that I do all that, then I am absolutely sure that I perform twice better; in other words a 100% increase in efficiency, avoiding syrupy bureaucratic procedures and avoiding an abundance of building mistakes, avoiding the production of waste material, and accidentally keeping the building site extremely clean. How sustainable can you get? It is obvious that the nonstandard architect, who controls the efficiency of the process, must be the first to take profit from that expertise.

Controlling the design process from concept, from the first sketch all the way down to CNC production and methods of dry assemblage, gives the designer also control over the costs of the whole enterprise. Controlling the costs from scratch means having a powerful weapon in hand to compete on the market with large contractors and developers. The direct link between parametric 3D model and execution/assemblage allows the designer to take on the role of the contracting party and developing party him/herself. In the ONL projects Cockpit and Acoustic Barrier for example, we have embarked on a design and build organizational model that allowed us to offer the buildings as a product for a fixed price, which was very competitive as compared with calculations based on standard procedures and cost estimations based on market prices. For us this has been living proof that a truly nonstandard architecture featuring complexity based on simple rules is competitive with respect to regular modernist boxes of similar quality. In our times of a serious recession in the traditional building industry we have found a way to deal with this crisis. Independently from the traditional powerhouses in the building industry we enjoy having proven the affordability and sustainability of the new kind of building.

The appropriate way to effectuate the new role of the architect is to take part in the building process financially. In the present situation architects leave the financial responsibility to the project developers and the contractors, the architects themselves acting as a consultant only without being responsible for more than their designer fee. I am an advocate of a new professional attitude of the architect, to become an entrepreneur in their own right, to take over the responsible role of the contractor such as for all components which are CNC produced. Architects are chicken if they do not have the guts to claim the leading role such as a responsible designer–engineer–builder.

Back to Top

Document information

Published on 12/05/17
Submitted on 12/05/17

Licence: Other

Document Score


Views 26
Recommendations 0

Share this document

claim authorship

Are you one of the authors of this document?