Rethinking Prototyping

Bridging the Gap

Daniel Lordick and Caroline Spliid Høgsbro

1 Background

Bridging the Gap was a course for the development of a pedestrian bridge over a stream in the Berlin Park Großer Tiergarten held at the Technische Universität Berlin (TU Berlin) during winter term 2011/12 and summer term 2012. The course introduced parametric modelling to architecture students who had no previous programming knowledge. This paper reports about the method we used to gain highly specified bridge designs.

Parametric modelling calls for an appropriate kind of design thinking. This is evident from an on-going debate on computational design (cf. Burry 2011, Carpo 2012 et al.). Instead of a conventional result-driven approach in architecture, parametric modelling requires a process-driven approach. We pushed this concept by conducting a specific sequence of exercises and tutorials. At the same time, we established the idea of prototyping in order to link the early phases directly to the potential techniques of manufacturing (cf. Adenauer; Petruschat 2012).

The course is based on experience from previous courses held at the TU Dresden. The focus is set on designs strategies that significantly reflect the utilization of the computer beyond the common virtualization of traditional drawing techniques. This concerns generative design and as a side effect altered the way we communicated about the projects. The latter more representative aspect is captured in [Lordick 2013].

Daniel Lordick and Caroline Spliid Høgsbro

TU Berlin, Germany

2 Introduction

Parametric modelling is the integration of programming techniques into the design process. This means both, the use of specific software and the evolution of new concepts. With parametric modelling the design team can manage highly complex design tasks with high precision, generate design variations in real time, review the project with extraordinary flexibility and speed, and directly trigger the industrial production routines towards custom manufacturing. This implies that former boundaries between the technical and the creative realm lose their meaning. Software like the plug-in Grasshopper for the CAD software Rhinoceros has speeded up this evolution (Grasshopper 2013).

To stress the full potential of the approach, we take a closer look and specify the expression parametric modelling by the three terms parametric, associative and generative. Strictly speaking, the term parametric only means that design decisions are transferred into changeable values, the parameters. This is fundamental and already covered by any object-oriented architectural design software (CAAD). What parametric wants to emphasise here, is: The designer can determine the way parameters are defined and how the parameter-driven components are connected and affect each other. This functionality is more precisely referred to by the term associative and reflects that virtual models, created with this premise, can easily adapt to changing constraints. The advantage of wisely structured associative models is that they can be manipulated by only a few parameters but are still highly flexible. The third term in connection with parametric modelling is generative design. This concept goes even a step further and implies emergence and simulation (Bonacker et al. 2009, p. 463; cf. Johnson 2001). Emergence means that the result of a parametric modelled structure cannot be predicted exactly by reviewing the elements of the program. This is typical for agent-based systems (swarm behaviour) and recursive programs (Fig. 1). Simulation refers to software models of physical phenomena. Dynamic relaxation for example can yield in forms, which are rather defined in a meta-level than by direct input. In summary, parametric modelling is not only modelling the pursued design but also the design process itself. The designer now is able to control the design as a whole and to any detail desired at any stage of the development. The elaboration of the parametric model is part of the progress.

Fig. 14 Models from the first semester of Bridging the Gap.

Fig. 1 Parameter controlled fractal branching structure without any random script components (DL 2006).

It is not an easy task to make architecture students become acquainted with the concepts and techniques of parametric modelling. One reason is the harsh beginning and a relative long phase, before the approach plays off its full potential (Fig. 2). Sufficient motivation can only be achieved by projects not too large for beginners but already too complex to be managed by traditional means. Thus, several courses previous to Bridging the Gap focussed on digital production (Fig. 3, left) or geometrically challenging tasks (Fig. 3, right).

Fig. 2 Diagram according to Neil Katz, SOM, as seen at the Design Modelling Symposium Berlin 2011.

Fig. 3 Left: non-rationalized surface model from a laser-cutter in a file-to-factory manner (Silke Maret, TU Dresden, 2009). Right: pre-rationalized surface discretized by interlocking elements (Robin Bongers, TU Dresden, 2010).

3 Objectives

Before starting Bridging the Gap, we analysed the results of the previous courses and decided to gear the course strictly to the accessible and affordable means of production. Given the equipment of TU Berlin laser cutting for the scale models and three-axis CNC milling for components of the bridge seemed to be the suitable tools. Thus we envisioned the bridge to be constructed modular with custom parts and joints from plywood. Through the restriction to a certain technical outcome we ensured that the students were able to track the workflow from the early design stage to the file-to-fabrication phase by one coherent digital model.

The approach involves a culture of thinking that accentuates the definition of the geometrical structure instead of the actual form (cf. Valena et al. 2011, Nake et al. 2007, p. 220). A parametric model in essence is only meaningful, provided it is wisely structured, allows the creation of substantially different variations from a relatively small number of parameters, and at the same time can easily adapt to changing constraints. Thus we wanted to review not only the outcome of the Grasshopper definitions but particularly the way they were conceived.

Pedestrian bridges in the Großer Tiergarten stand for the integration of artefacts into a natural environment (Fig. 4). We wanted to refer to this topic by looking at nature at a micro scale instead of mimicking its phenomena. Conceptually, there is an interesting coherence between the growth of plants from genome information and the creation of parametric models from bits of code (cf. the term digital morphogenesis, e.g. Hensel, Menges 2008). Although a species as well as a parametric model is essentially described by its code (genotype), there is still a great variety of forms possible by slightly changing the variables and by inherent adaptability (phenotypes).

Fig. 4 Detail of a historic pedestrian bridge in the Tiergarten with a floral motif.

One pitfall of parametric modelling is that the students forget about their ability to design and instead clone code solutions found in the Internet – with the effect of a worldwide homogeneous menu of inchoate parametric compositions at any digital design lab (van Berkel 2013). One reason for this apparent defect may be that the students are simply not able to take advantage of the programming tools because of a lack of deeper understanding. Thus, the programming environment again – as any other CAAD software – acts as a black box with predefined tool sets and outcome. The second reason might be found in the software itself: Only design decisions that can easily be quantified find their way into the digital model. Last but not least the students are often overwhelmed by the efficiency of parametric models in handling large quantities, which may suggest complexity. It takes a little experience to realize that apparent complexity is not added value.

The challenge in conducting Bridging the Gap was to avoid these tendencies by two strategies. First, in order not to be narrowed by the software the students evolved the seedlings of the designs with analogue techniques. Second, the students had to build up Grasshopper definitions from scratch. In a set of consecutive tutorials, each starting from a plain canvas the students became familiar with the essential concepts of the software and the underlying geometric knowledge. Altogether the aim of the course was to unleash the students’ potential to handle complex design tasks skilfully and to make them create unique concepts within a short period of time.

4 Method

In respect to the location in the park and with the conceptual orientation towards a modular structure we started our course with the analysis of the formation and structure of elements and joints from biological examples (Fig. 5). Each student had to work on sketches and physical models to document personal observations and speculations (Fig. 6). The students investigated in several directions: Is an example promising from the structural point of view? Does it reveal geometrical principles that qualify it for a module? Could it be an inspiration for materialization in a larger scale?

Fig. 5 Some found objects serving as examples at the beginning of Bridging the Gap.

The reflection on biological artefacts has numerous famous precursors like Karl Blossfeldt and Ernst Haeckel (Sachsse 1996, Haeckel 1998). We especially emphasized the morphological perspective that D’Arcy Wentworth Thompson unfolded (Thompson 1917). Though, the focus was not to generate engineering solutions like in bionics, but to start a voyage of discovery as for example artist Amely Spötzl undertakes in disassembling and cutting parts of plants (Hupasch; Lordick 2008, pp. 24/25, 70/71). During this first step, each student unveiled a pattern and principle to refer to during the following tasks. Then groups of two students each were formed to combine related motives. In cooperation the students formulated a spectrum of forms that transcended the biological example, and they tested and extracted options of parameterization for the digital model (Fig. 7). Next the concepts were transferred into code and the resulting digital models were refined. In a phase of exploration and inspection the students created variants gauging the design space of their concepts (Fig. 8).

Fig. 6 Selection of students’ sketches and models inspired by biological examples.

Fig. 7 Formal explorations to extract relevant parameters.

Fig. 8 Variations derived from a parametric model.

Obviously, we established a procedure where we did not start with the instruction „Design a footbridge!”, but in a bottom-up strategy tried to develop and agglomerate building parts, which eventually are able to span a stream. This is a generative method, which among other aspects helped to prevent unfounded copy and paste tactics during the scripting phase: Any new inspiration appearing on the screens had to be related to previous steps of the evolution.

Generative design is not an invention of the digital era. It is a highly process-based approach that opens the space of possible solutions for the unexpected. In a systematic sequence of actions that are partly carried out by technical means. Step by step the design emerges. The idea is not to draw from a source of predefined forms but to get involved into an algorithm for form-finding (Kraft; Taraz-Breinholt 2002, p. 20). As an early predecessor for generative strategies may apply Sonia Landy Sheridan, who in the field of fine arts experimented with photocopiers and fax machines during the 1970s (Kirkpatrick et al. 2009). The digital tools we use today are certainly much more sophisticated. As far as parametric modelling is concerned, the most remarkable feature is that at any time the whole process from the initial assumptions to the ultimate fabrication data can be reviewed and controlled interactively.

Bridging the Gap had the aim to create a pedestrian bridge. This is a task typically carried out by civil engineers alone. We wanted to question and overcome this by linking radical structural and formal thinking. During the first semester we had two invited talks contributing to this topic. The first was by Lorenz Lachauer, who introduced his form-finding tool for curved bridges (Lachauer; Kotnik 2011). The second talk was by Prof. Mike Schlaich, who reported about his essentials and about exemplary projects from his long practice in designing footbridges (Baus; Schlaich 2013).

Fig. 9 Three selected projects from the first semester of Bridging the Gap.

At the beginning of the second semester we selected three objects that had to be developed further by three groups of students (Fig. 9). The aim was to explore the bridges in terms of their potential for actually getting manufactured and to inspect their inherent structural behaviour (Fig. 10). Finally, one design was rated to be most promising and selected for in-depth examination. Now the students collaboratively edited the details and created parts with the milling machine. Three significant units of larger joints were assembled. The units were then stress tested in collaboration with the department of Prof. Mike Schlaich (Fig. 11.).

Fig. 10 Studies accompanying the selection process.

Fig. 11 Component, stress test of a standard unit, assembled elements.

The communication about the progress during the course was a main issue. However, the traditional repertoire of floor plans, elevations and sections once claimed by Leon Battista Alberti (1404-1472) did only partly apply to the rather complex shapes and the dynamically changing designs. For this reason we supported additional methods of visual representations to broaden the graphical repertoire. The students explored flow charts, screen shots of the visual programming software on the fly, sequences, and design space diagrams (Fig. 12).

Fig. 12 Flow chart documenting the process of a student’s project.

5 Outcome

In a research-by-design mode, the course Bridging the Gap explored the potentials of parametric modelling for a small project. At the same time it was a case study for the question how the method of generative design can be taught. This was answered with a process driven approach. The students were engaged with inspiration from biology, created their own personal genotypes in Grasshopper and in the end, nature is still present in the designs, not in form of mimicry but as a principle. They learned to generate forms according to their principles that they could not have imagined in the beginning.

Although they were freshmen, all students in Bridging the Gap really got involved with parametric modelling and pursued it to the final presentation. We were able to observe that they valued the new capabilities and entered regions that would have been unachievable without parametric tools (Fig. 13). The evident outcome is a number of models that are both, mysterious and inspiring (Fig. 14.).

Furthermore, the students contributed convincing examples to an evolving language of parametric design by exploiting and combining diagrammatic forms of representation. The depiction of sequences, design spaces and flow charts corresponded to the process of design decisions (Fig. 15). In fact, the design teams realized that for their ability to cooperate, it is indispensable to continuously broaden the visual vocabulary (cf. Lordick 2013)]

Parametric modelling has the potential to integrate substantial simulations of loading conditions and can hence foster the collaboration of civil engineers and architects. One of the participants will further elaborate this aspect in the course of his bachelor’s thesis.

Fig. 13 Joint work in the second semester of Bridging the Gap (Anastasia Vitusevych, Tobias Kuhlmann, Georgios Chousen et al., rendering: Benjamin Weichert).

6 Conclusion

The major advantages of parametric models are flexibility and adaptability. In order to harvest those benefits to their full extents, a parametric model has to be set up consistently from the very beginning up to the ultimate assembling of the project. It is not sufficient to retrieve a design that looks as if it could have been done by generative methods. Instead, the code must have the potential to spread a certain range of different designs. This means that the supervisors need to in-depth review the students’ approach to avoid later frustration. We are not aiming at the creation of specific shapes but at the awareness of the potential of thinking in structures and correlations.

During the course we learned that we lost a lot of precious time for technical instructions. As an answer to this educational question, we created a basic course on parametric modelling for the first-year students. The objective of this new module is to provide the essentials of the tools and let the students experience a typical workflow from the definition of a simple model to its production with a laser cutter. This shall be the first seed for the integration of parametric modelling knowledge into design thinking.

We do not believe that every design task should be carried out by the means of parametric modelling, but we want to equip the students with the knowledge needed to exploit the benefits of parametric modelling whenever useful. Parametric modelling is more than the integration of just another tool. It affects the designing culture and the way we communicate about our projects.

Fig. 15 Example for a presentation that is emphasising the design process.

References

Adenauer, J.; Petruschat, J., 2012: Prototype! Physical, Virtual, Hybrid, Smart. Tackling New Challenges in Design and Engineering. Berlin: form + zweck.

Baus U.; Schlaich M., 2013: Footbridges: Construction, Design, History. Basel: Birkhäuser.

Bonacker, H.; Groß, B.; Laub, J., Lazzeroni, C., 2010: Generative Gestaltung. Entwerfen, Programmieren, Visualisieren. Mainz: Herrmann Schmidt.

Burry, M., 2011: Scripting Cultures: Architectural Design and Programming (Architectural Design Primer). New York, NY: Wiley.

Carpo, M. (ed.), 2012: The Digital Turn in Architecture 1992-2012. New York, NY: Wiley.

Grasshopper 2013: Website of the Plug-In Grasshopper for the CAD Software Rhinoceros: http://www.grasshopper3d.com [05.05.2013].

Haeckel, E., 1998: Kunstformen der Natur. München; New York: Prestel.

Hensel, M.; Menges, A. (eds.), 2008: Designing Morpho-Ecologies: Versatility and Vicissitude of Heterogeneous Space, Architectural Design. New York, NY: Wiley.

Hupasch, V.; Lordick, D., 2008: Good Vibrations – Geometrie und Kunst. Dresden: Univ.-Sammlungen. Kunst + Technik.

Johnson, S., 2001: Emergence. The Connected Lives of Ants, Brains, Cities, and Software. New York: Scribner.

Kirkpatrick, D.; Flanagan, M.; Hart, K., 2009: The Art of Sonia Landy Sheridan. Hanover, NH: Hood Museum of Art.

Kraft, S.; Taraz-Breinholt, S., 2002: Formfindungen von biomorph bis technoform. Arch+ 159/160, p. 20.

Lachauer, L.; Kotnik, T., 2011: Curved Bridge Design. In: C. Gengnagel, A. Kilian, N. Palz, F. Scheurer (eds.): Computational Design Modeling, Proceedings of the Design Modeling Symposium Berlin 2011, Berlin Vienna: Springer, pp. 145-152.

Lordick, D., 2013: A Visual Language for Parametric Modelling. Research in Architecture. Berlin: Universitätsverlag der Technischen Universität Berlin, pp. 162-165.

Nake, F.; Grabowski; S., Hashagen, A.; Krauß, M., 2007: Angeregt von Paul Klee und Manfred Mohr. Bewegungen im Hyperwürfel und auf der Bildfläche. In: Niewiadomski, O. (ed): Geometrie, Kunst und Wissenschaft. Bremen: Hauschild.

Sachsse, R., 1996: Photographs 1865 - 1932 / Karl Blossfeldt. Köln; Lisboa; London; New York; Paris; Tokyo: Taschen.

Thompson, D. W., 1917: On Growth and Form. Cambridge: University Press.

Valena, T.; Avermaete, T.; Vrachliotis, G. (eds) 2011: Structuralism Reloaded. Rule-Based Design in Architecture and Urbanism. Stuttgart; London: Edition Axel Menges.

van Berkel, B., 2013: Dean’s Statement: http://www.staedelschule.de/architecture/dean_pages_ben_van_berkel.html#c114 [04.04.2103].