Let your Computer be a Designer

Scientific breakthroughs and technological advancements are transforming urban planning from a product-oriented artistic endeavor towards a process-oriented scientific discipline.

As the focus is shifting from what we design to how we design, we need a more experimental approach towards new technology. Digital tools allow planners to think and act on a fundamentally larger scale. However, they also require the ability to express our ideas and objectives using computational language.

The genius is the talent that gives rule to art. High art, consequently, cannot be born from anything else than from the mind of the genius.

Immanuel Kant, Kritik der Urteilskraft (1790)

Before the 15th century, creativity was considered a divine quality. Not until the Italian Renaissance did the notion grow that remarkably talented painters, sculptors and architects were capable of creating new worlds out of nothing by simply using their creativity.

In 1790, German philosopher Immanuel Kant conceptualized the term genius in Kritik der Urteilskraft. Kant made a radical distinction between the aesthetic genius and the scientific mind: the genius was someone who could produce ideas without imitating nature or learning from others.

Kant’s theory has been rooted in our cultural tradition ever since. Design, literature, fine arts and almost any other form of creative output typically relies on the individual genius’s imagination as its major source of input. Or this is how it is mostly portrayed in autobiographies, monographs and other commemorations that all celebrate the individual creatives as the foundation of our culture.

The model of the genius has had a fundamental role in architecture and urban planning, as well. The theory was particularly favored by early 20th century modernists eager to conceptualize the creative processes that take place in an architect’s mind. For instance, the famous doctrines ‘Less is more’ and ‘God is in the details’ by German architect Ludwig Mies van der Rohe bear close resemblance to Kant’s conception. They present the architect as a talent who is able to harmonize and simplify the numerous components of a building.

The majority of our digital tools and techniques are also built upon the traditional conception of creativity. Although the work itself is—to a growing extent—made using computers, the elements that constitute a design, plan or other composition are typically positioned manually by a human (or a group of people). Most designers tend to regard computers as mere visualization tools and are unwilling to investigate alternative ways of utilizing them during the creative process—unless you made it from scratch, you did not design it!

Urban planning is an information-based process

All processes, whether they are produced by human effort or occur spontaneously in nature, can be viewed as computations.

Stephen Wolfram, A New Kind of Science (2002)

The Principle of Computational Equivalence was introduced by English mathematician and scientist Stephen Wolfram in 2002. The theory suggests that all mechanisms, whether cultural or natural, can be regarded as information-based processes controlled by the same underlying mathematical logic. Thanks to this logic, all these processes can be reproduced and examined using computational technology.

The deeper we understand the laws of nature, the more our perception of the physical space changes. The city, which has been portrayed as an arrangement of buildings and other static objects, is transforming bit by bit into a system of dynamic relationships and flows of information. Since urban planning has traditionally been conceived as a discipline that concentrates on shaping a city, it is apparent that the nature of planning will change as we gain more insight into the laws of the city.

Many laws of nature, such as gravity and collisions between two or more objects can be modeled (in their simplest form) with mathematical equations and a few lines of code.

In addition to the information-based approach there is another, even more distinctive, trend changing the course of the discipline. Computers—which were introduced in the 60’s and 70’s as an alternative means of conducting drafting—are today replacing almost all previously used design, drawing, modeling, visualization and communication techniques.

Still, many architects and urban planners neglect the computational approach and utilize computers like 50 years ago; for creating more detailed drawings and illustrations. This is largely due to our current built-in conceptions of creativity and creative work. Architecture and urban planning are considered artistic disciplines, in which the creative process takes place in the designer’s mind and not on the microchip of a computer. However—as Wolfram’s principle suggests—this process can also be broken down into computational models.

Computers think faster than the designer

Like a pet, the computer has already been domesticated and pedigreed, yet it does not behave with human intelligence. Just as a pet introduces an element of wildness to our domestic habits that must be controlled and disciplined, the computer brings both a degree of discipline and unanticipated behavior to the design process. By negotiating the degree of discipline and wildness, one can cultivate an intuition into the behavior of cad systems and the mathematics behind them.

Greg Lynn, Animate Form (1999)

In a typical design process, the designer makes sequential decisions which originate either from pure intuition or from more explicit factors, such as planning regulations, economical constraints or statistical facts. Converting this process into a computational model means that the decisions and other steps taken in the designer’s process are expressed as explicit rules, parameters and functions, which can be remodeled using computers.

Kaisersrot was a research project at ETH Zürich led by architect and computer scientist Ludger Hoverstadt. The project explored a bottom-up approach to computer-aided urban planning and resulted in a series of applications that utilize agent-based simulations to find solutions to various design problems.

One of the most interesting results is an application that is presented in the video Follow the church. In the application, the individual plots are represented by agents that are assigned certain rules and constraints. The agents are aware of factors such as their building type and size, their preferred neighbors and their desired proximity to water, parks, shops and other external attractors.

Agent-based approach to urban-planning, by Kaisersrot.

The agents start from random locations and are then allowed to push and pull each other until they find a position that suits their built-in preferences. The overall plan for the area emerges in less than a minute through the agent interactions, without any assistance or intervention by the user.

It is important to notice, though, that the computer itself is not deciding what to design. It simply does what a computer is best at doing: very quickly working out the consequences of the constraints and criteria defined by the user. This enables the planner to create an unlimited amount of valid proposals and manually work further on the most interesting ones.

The whole process differs fundamentally from classic top-down planning. The application simultaneously provides a graphical visualization as well as an in-depth analysis of different criteria like costs, infrastructure or housing types. And all this is achieved in a fraction of the time required by typical means. The computer acts as an extension to the human brain, not as a more accurate pen.

Urban planning meets agent-based modeling. Predefined geometrical constraints and a few interaction rules result in an unlimited number of valid plan drafts (which must be completed by a designer, of course).

From the Aesthetic Genius to the Scientific Mind

The architect who proposes to run with technology knows that he will be in fast company, and that, in order to keep up, he may have to emulate the futurists and discard his whole cultural load, including the professional garments by which he is recognised as an architect.

Reyner Banham, Theory and Design in the First Machine Age (1960)

The computational approach is not tied to any specific program or scripting language. It represents a general way of utilizing the computers’ information processing capacity during various stages of the planning process. Computational techniques do not differ much from trial and error based manual procedures—the computer only does the job much faster and more efficiently than a human.

Besides boosting the creative processes, digital tools can be used to improve communication between the planners and various stakeholders. Computational techniques allow the comparison of alternatives on a rate that is inaccessible when working with traditional design methods under traditional time constraints. They also enable planning beyond the common top-down approach, when plans can be derived from parameters that leave room for negotiation.

It is apparent that scientific breakthroughs and technological advancements are transforming urban planning from a product-oriented artistic endeavor towards a process-oriented scientific discipline. This puts a growing emphasis on digital tools that allow us to think and act on a fundamentally larger scale. Perhaps even more importantly, it calls for slight adjustments to how we see ourselves as architects and urban planners. Instead of being the authors of drafts, we have become administrators in a process: Great cities cannot be born from anything else than the scientific mind.


To hire me, request an appearance or discuss a project, please contact me via jyrki@jyrkivanamo.com, +358 50 547 0020 or LinkedIn.

With kind regards,

Signature of Jyrki Vanamo