(originally written for M.S. degree theory class at UMICH)
The tool’s conceit
Architects, draftsmen, engineers, scientists, illustrators, and professionals of many trades have used the French curve. When they draw, some lines are easy to construct while others are difficult to capture with an easy, prescriptive method. These difficult curves are often approximated with what is commonly referred to as a set of French curves. The role it plays calls for examination, particularly at a time when architecture’s expertise is shifting towards technicality under the aid of new digital tools not unlike series of “new” French curves.
First of all the French curve as a tool and its seemingly organic evolution may have been an effete attempt to advance precision in craftsmanship after all, and further unveil a peculiar attitude of us, namely designers whose professional obligation is visual representation by drawing or drafting. This attitude clings onto precision, dangerously defaulting design responsibilities.
The known beginning of the French curves was in the 1800s. They are sets of templates that contain many segments of an Euler spiral, which by definition tightens its curvature as its length grows. With a set of differential equations, an Euler spiral is believed to encompass all curvatures possible. By incorporating these curvature into French curves, Man has the tool to draw difficult curves. With the French curve as a present-at-hand tool, in Martin Heidegger’s term, one can declare that Man can draw all curves.
Tool as crutches
French curves’ mathematical perfection is only transferable as an epistemic description. A few equations give the Euler spiral. Cut segments from it and splice them together, and we have the tool. In practice, segments of other mathematical curvature can also emulate difficult curves. Even a set of indistinct or seemingly random curves may come to approximate a desired line. We are naïve to prefer a tool’s applicability by its concepts, or even its conceits.
Somehow French curve’s own definitive precision legitimizes lines it produces. This tool becomes a crutch. Philip Johnson has warned us of the crutches in architecture. To a great extent, the French curve is the second kind of crutch Johnson mentioned. A French curve to pretty drawings is a pretty drawing to architecture. Pretty drawings are not equivalent to great architecture. Likewise the French curve is of little importance to the making of curve lines.
Architects have long moved beyond the French curves. It has been observed that since the mid-1990s, architects’ infatuation with Non-uniform rational B-spline (NURBS) modeling opened the flood gate for many new digital tools, many of which comes with unfathomable precision. Design soon witnessed its new born child, computation – a new French curve of our time.
Every designer equipped with the knowledge of design computation cannot stop cranking out swarms, flocks, and armies of complex shapes dictated by numbers and proportions. Illustrated here is a tool called Culebra. Ironically what many describe as a certain topological random walk is precisely determined by a number of points, locations of motion origin, displacement vectors, frame rate, etc. Random turns out to be just designer’s default. Perhaps many of these designers look on, agape at the void of meaning these visually “loud” objects leave behind. They are conceived with such prescription and precision that they must be truthful, as we tend to think. Somehow in our head a voice echoes: ‘Look! It is staying true to geometric rules’. If it’s truthful it must be meaningful. Yet it remains difficult to pinpoint what these wrapping white wires with inscrutable paths are doing for us.
In her book The Architecture of Error: Matter, Measure, and the Misadventure of Precision, Francesca Hughes unfolds our reliance on precision, sometimes a reliance in vain. She elaborates on the causes of our precision fever with two accounts. One refers to our expanded horizon in natural sciences since the 18th century. The other is our modern culture of standardization plus the intolerance for derivation. More specifically, architects’ fetish may come from the professional pretension to science. Design may connote an unerring method, full control. Soon, the demand for precision in production devolved into an administrative norm. We are often asked to achieve great resolution, no longer driven by the push-back on errors. As it serves political agendas, precision begins officiating design.
Moreover, computers produce precision so easily that it often exceed the resolution of both of its source data and the target product. In the virtual model space, does the virtue of a tolerance down to the 1/10 of a millimeter extend beyond its legibility by a robot? The precise representation may never be enforced beyond paper. “The quantum increase in precision that has accompanied the digit[ization] of architectural construction far outstrips the equivalent increase that shadowed the rise of industrialized manufacturing and quality control driven by two world wars.” (Hughes, p19). As far as drawings for construction go, this fidelity to integrity often passes unnoticed. On site, it is a world of shims and caulks.
One fabrication method seem to match the level of precision ubiquitous on our digital landscape - digital fabrication. The robotic arms, the laser cutters, the fused deposition modeling gadgets (FDM 3D printing) all promise the designer instant translation from their computer screens to the material world. However, these high-tech contrivances cannot alter material properties.
For example, the might of the robotic arm wanes at the tip of its end effector. Location 1 is where most of the calculation occur. The computer monitors every coordinates to the 3rd or more decimal place. Location 2 is the robot itself, which reads strings of numeric values without a hiccup. Step 3, the end effector affects the material. Be it a nozzle extruding molten plastic or a red hot metal wire slicing through foam, errors emerge. The cooling molten plastic tends to recoil while the flexible hot wire bows in the inconsistent density of the block foam. True accuracy is only discovered by experience. Through repetition, one approximates the “precise” outcome, not unlike a seasoned craftsman drawing the perfect line un-aided after much stroke practice.
Jumping scales is another nuisance that chokes design precision. The larger the structure, the more evident expansion/shrinking in materials become. As the number of unit repetition increases, errors in building parts accumulate. With the massive computing power current technology allows, these unpredictability may be tackled by finite analysis and behavior simulations. However, that is the moment when computation or the broader digital technology is no longer a tool for a designer, but machine set up in an engineer’s workspace.
“Computation miraculously procures exactitude with more ease than approximation. Approximation is where the work starts”, wrote Hughes. She then brought in the incisive quote from philosopher Nancy Cartwright: “…every theory…required regularity… [But] there are no genuine regularities at the phenomenological level. Approximations take us away from theory and each step away from theory moves closer towards the truth.” (Hughes, p30)
Neither the French curve nor design computation, nor any form of crutches for this matter, validates a design. For accuracy in practice, the profession relies on gains through iterative trials.
Machine over tools
As far as drawing lines are concerned, the best line represents the subject, independent of how good the tool is. As an alternative to the French curve tool, we need a (hypothetical) plotting machine that gives us the actual points on a line. If for a moment we idealize this machine, it can produce dots so dense that they effectively become the line. That is the precise one.
A machine repetitively produces predicted results. It is vernacular to industrial production much more than tools. Machinery lives in a factory. A tool lives in the hands. Rather than searching for precision in tools, the designer benefit from machine’s efficacy. If an architecture’s value resides in the design of buildings, all the data, numbers, sequences, or strings should inform it, not so much dictating.
Recognizing the instrumentality of digital tools, designers today must learn to treat them as machine. They serves to automate, expedite, and optimize. It is false to assume that with the right tool one could be omnipotent. But with the right machine, one’s skills are complemented, productivity enhanced.
When a tool drives design or starts to harness a formal, even stylistic thesis, the architect must remember to “walk” again without the “crutches”. When an architect marvels at the visualization of algorithms, s/he must recount his/her job as more than diagraming the abstract. As the founders of UN Studio write in one of their essays, architecture profession is in the biggest network it has ever been. In face of “plug-in professionalism”, the architect will need help with some tasks while they manage others. Architects must know their role. Enhanced precision doesn't design.
Bibliography
Christian, Brian. The Most Human Human. Doubleday. New York. 2011
Crary, Jonathan. Techniques of the Observer: on Vision and Modernity in the 19th Century. MIT Press. Cambridge, Mass. 1990.
Hale, Jonathan. Harman on Heidegger: ‘Buildings as Tool-Beings’. https://bodyoftheory.com/2013/05/29/harman-on-heidegger-buildings-as-tool-beings/
Heidegger, Martin. Being and Time. Harper & Row. New York. 1962.
Hill, Jonathan. Drawing Research. The Journal of Architecture, 11:3, p329-p333. June 2006.
Hughes, Francesca. The Architecture of Error. The MIT Press. London, England. 2014
Johnson, Philip. Seven Crutches of Modern Architecture. Perspecta, Vol.3 p40-45. 1995
Nowacki, Horst and Lefevre, Wolfgang. Creating Shapes in Civil and Naval Architecture: a Cross-disciplinary Comparison. Koninklijke Brill NV. Leiden, the Netherlands. 2009
Schwartz, Hillel. Second Nature. The Culture of the Copy: Striking Likenesses, Unreasonable Facsimiles. Zone Books. New York. 1996
Van Berkel, Ben and Bos, Caroline. The New Concept of the Architect. MOVE. Goose Press. The Netherlands. 1999.