A render is a promise

A convincing image of a building comes with many claims that you can’t see. The roof can stand, water will go somewhere, people can reach the entrance, somebody can clean the glass, the budget has something to do with the shape. The image itself establishes none of them.

I’m not against speculative images. An impossible building can be a great way to think, and fiction lets us look at alternatives before feasibility ends the conversation. The trouble starts when a proposal looks so finished that people take the look as evidence that it’s ready.

What helps me is to split the question: - Can I describe it coherently? - Can it work? - Can the people it’s meant for get it and use it? - Can I defend its consequences?

A good answer to one says little about the others. A design can be structurally fine and still out of reach for the people it’s meant for, or affordable only because nobody counted the maintenance.

Luciano Floridi’s paper The Logic of Design looks at design through requirements and the systems that can satisfy them. The practical lesson I took from it: before asking whether a proposal is impressive, write down its requirements.

A physical example I like is the Smart Slab from ETH Zurich. In 2018 they made a concrete floor structure using 3D-printed sand formwork. The concrete itself wasn’t printed. The research had to connect computational geometry with molds, fabrication, assembly and structural demands, and that whole chain is why the project tells you more than a picture of a similarly complex ceiling would.

In software, the same gap is between a demo and a service. A demo gets ideal input, a cooperative network and a developer ready to step in. A service gets users who misunderstand it and failures that come at the worst possible time. Both are fine, they just answer different questions.

Simulation is the case closest to my own work, since simulation software (Siml.ai) is what we build at DimensionLab. A simulation supports a claim within its assumptions, and it’s honest only if you say what those assumptions are.

As generated images get cheaper, I expect that saying clearly what a representation proves will become a more valuable skill. So the next time something looks inevitable because it looks finished, I want to slow down and ask which parts of it have already met something that can push back.