The sectional drawing has all but disappeared as a way of conceiving, depicting and communicating architectural space. I think the rot set in with those CAD programs students like to believe “do it all for you” and that all one has to do to generate a section at any time is slice the model along a cutting plane. Other than it looking something like a section, the point of doing this isn’t understood. This mindset assumes the purpose of the section is to verify what’s going on but why do that when a flythrough of a fully modeled building can show you better? The greater issue is that the more visualization that happens onscreen, the less that happens inside the designer’s head where the real virtual model should exist. Students and instructors should be wary of over-reliance on such tools.

A similar atrophication of architectural skills is happening with models where the ideal seems to be making an accurate one with as little effort as possible and without any of the learning about size and configuration of spaces that are supposed to be the purpose of making a model. The readymade high-resolution model is what the marketing and exhibition industries prefer. I don’t blame students. It’s merely a reflection of the industry wish for scaled replicas for promotional purposes. 3D printing accelerates this dumbing down as any hope of feedback between thought and result is lost, along with any intimation of materiality and construction and, more to the point, the paid labour that goes into the making of real buildings. Once again, the academic world unthinkingly tracks the real one. We’re not completely there yet but the stage is being set. Today’s dominant aesthetic is one of Shape in which no trace of materials or labour remains. [Peter Cook, when reviewing ZHA’s Aliyev Center was made of, famously omitted to mention what it was made of. I see this not as an oversight but an example of educators following whichever way the wind blows.] The 3D printed model of an architectural idea exhibited as art is not a representation of this societal force. Divorced from even the reality of buildings, it’s its perfect embodiment.

Before admitting defeat and giving up on the section as having anything to offer architecture education, I thought I’d give it one last try. It goes like this.

1. We’re already familiar with CT (CAT) scanners that take a series of vertical cross-sectional x-rays, allowing doctors to understand the state of our bodies.
2. We’re also familiar with 3D printing that creates 3D objects by depositing some arbitrary substance one horizontal layer at a time.

The goal is to construct a three-dimensional model of a house by gluing together vertical cross sections laser cut from wood.

This means a CAD program will be used to mimic the output of a CAT scanner and feed it to a laser cutter will be used to mimic the single layer output of a 3D printer. People will also mimic a 3D printer by gluing those individual layers together one layer at a time. In the studio, bouncing images between snazzy applications that can only do one thing is the norm and, sadly, is often taken to represent skill. Students learn to say things like “I did this in A and then exported to B, and then into C before back into A so I could ….”], so there’s something appealingly subversive about using known technologies in ways for which they weren’t designed. The important thing is for students to understand a section as nothing more or less than a different way of conceiving, comprehending and communicating the reality of a building.

STEP 1: Collect the Data

Obtain dimensioned plan, elevation and, if possible, section information for the building to be modeled. The building of my demonstration project will be Kazuo Shinohara’s 1976 House in Uehara for which I have sufficiently dimensioned plans and sections.

STEP 2: Draw the sections

Use the section data as the baseline, checking it against the elevation.

My sectional “scans” will start at the rear of the house and end at the front. House in Uehara is 9090 mm deep. A ply thickness of 2 mm would mean 4,545 sections would need to be drawn and cut at 1:1, 45 if the scale were 1:100, and 92 if the scale were 1:50. 1:50 seems manageable, and means the dimensions in plan of the final model will be about 45 cm x 45 cm, but it also means that a section will be taken through the building every 10 cm. This resolution may turn out to not be high enough to model the building to a sufficient fidelity but will probably be adequate for teaching purposes.

A ply thickness of 1mm, and not considering the thickness of the adhesive, would mean 184 sections for the same 1:50 model. This is twice the cutting and twice the gluing and still not particularly onerous but it would mean a resolution of 5cm which would be better to represent detail such as window frames. The thickness of the adhesive may begin to matter if it didn’t already. A test is needed to find out.

Creating these section “scans” is the most laborious part of the project but is also where the learning bits are. A dwg file containing red lines to be cut and blue lines to be “etched” is sufficient for output to the laser cutter. Onscreen however, those lines could represent “slabs” so the cumulative build-up of the 3D model can be monitored and checked as it proceeds.

Note 2: Some students will prefer to completely model the building and then generate sections from section lines placed every 10 or 5cm. These could then be transferred to the dwg layouts accepted by laser printers but they’d first need to be checked and any spurious or unnecessary lines removed. This itself would be a learning exercise, but all that would be learned is that the program can “do it for you” only if you already know what you’re doing. Moreover, the advantage of the 3D feedback described in Note 1 would be lost.

Note 3: A decision needs to be made regarding the depiction of glass and doors. Is the physical reality of glass going to be depicted as a smooth solid surface, or is the visual one going to be depicted by some transparent material? (The Greeks and the Romans had different approaches to the depiction of eyes on statues.) With House in Uehara, a decision needs to made on whether doors and ventilation panels are going to be openable. These questions are secondary to the point of the exercise, but worth thinking about and an approach settled on.

STEP 3: Laser Cut, Assemble

Laser-cut and assemble. Any two adjacent sections could be left without adhesive, producing a model that can be split into two pieces like San Carlino.

I know an architect gifted in spatial imagination. He draws and has always been a prolific sketcher whether doodling for fun or thinking things through with a pencil a way of understanding things. With just a pencil and paper he could instruct a team or an office. The virtual model of the building exists in his head. He’s proficient in various CAD packages but AutoCAD has a special place in his heart. He likens it to a pencil. “In order for it to do anything you have to push it.” This suggests that architectural imagination is not hindered by clunky tools such as pencils and AutoCAD but actually develops and becomes stronger because of them. In other words, these tools are sufficient. I’d go so far as to say that the clunkier the tool, the more likely it is to develop imagination and spatial thinking.

The corollary is that the more complex tools actually hinder, if not actively prevent the development of arhitectural imagination. Consider the exercise I’ve just described. Imagine if instead of Shinohara’s boxy House in Uehara, we have some onscreen shape being stretched, morphed, extruded, and walked and flown through until the result is a set of some interconnected spaces within some apparently amorphous mass and then select 3DPrint. We would still not get Frederick Keisler’s 1956 Endless House.

Somewhere along the way, we’ve constrained our thinking to suit the limits of the new applications we keep getting given, and every new one constrains it even more.

---

3D Printers in History

Architecture Myths #32: Representation