Representation shapes how we see the world, but it also limits what we see. This has profound implications for architecture, where our models of context dictate the buildings we create.
To understand the context of a given architectural situation, we model it using some kind of abstraction or representation. It is necessary to model the reality - both the built environment and abstract aspects of society - to test how our designs work with it. This can be in form of drawings, or more abstract forms such as text or art pieces. But our way of representing the context becomes our way of understanding it.
In computer science, a ‘leaking abstraction’ refers to when a model simplifies reality so much that it fails to fully capture it, and this failure—like water leaking from a pipe—reveals itself over time in unexpected problems. Architectural representations, like drawings or zoning plans, face similar challenges: they simplify the messy complexity of reality, but those simplifications often ‘leak’—missing key aspects of human life, history, or physical context.
Let’s say we describe the context of our project using neat and exact line drawings of the neighbouring facades. These are excellent for understanding form and proportion. Such a representation will most likely influence us to do a project exploring the proportions and forms of the facade in relation to their neighbours. But it also “leaks”— it misses the weathering, repairs, and messiness that give a street its character.
Photogrammetric meshes offer a fascinating alternative. These representations, created through data-driven machine processes, dissolve the city into a fragmented geometric soup. Unlike line drawings, they don’t prioritize human perception but instead show an alien and messy perspective on the city. These leaks are glaring—every surface is distorted. Yet, they are also illuminating, because they break our preconception of a perfectly ordered city.
Just as line drawings reduce a street’s lived character to neat proportions and photogrammetry distills it into rough, fragmented surfaces, municipal planning reduces the city to measurable metrics. Although it aims to account for all kinds of things, it’s outputs reveal the dominating abstractions: efficiency, cost control, land use optimization, and risk management. Needless to say these representational modes has severe leaks, failing to take into account what the city needs to be for humans to thrive.
In our project, we tried to understand context as a series of stories about people - a web of narratives if you like - to reveal a human side of the neighbourhood. It made us aware of the cholera hospital of the 1830s, the queer safe spaces of the 90s and the synagogue of the 40s. It opened our eyes to the conflicts between vice and virtue, with taverns and the sex industry living next door to centers of faith and political organization. According to this representation, the purpose of this neigbourhood can be understood as housing the marginalized, and giving space for the coexistence of contradictory elements.
We also looked at the change in the physical environment over time. The changing needs of urban life has led buildings to change again and again, away from their intended needs. A productive conflict between the strict urban grid and many small ad-hoc interventions has created character and interesting inconsistencies—a quality Robert Venturi described as ‘messy vitality,’ where disorder becomes a source of richness and creativity
What, then, should the purpose of our project be? Our investigation led us to a conclusion: this must be a space for care and inclusion—a space that brings the poor alongside the rich, resists the imposition of sterile urban order, and embraces architectural contrast and “messy vitality.” Therefore, we propose a homeless shelter, tucked between two streets, accessible through a public rose garden tended by the shelter’s residents. It insists that this land can be reimagined as a place of beauty, care, and dignity for the marginalized.
Is it utopian? Perhaps. But is it any more naive than business-as-usual? Cities need spaces for beauty, care, and dignity. Ignoring these needs has immense costs, even if they are not modeled in the representational modes which dominate. Our project embraces the leaks—where abstract systems fail—and imagines a space where care and complexity can coexist.”