Forward to basics

It doesn’t often happen that architecture, including the process of design, construction and use, is given a fundamental overhaul. In his recently published book Towards a New Kind of Building, however, Kas Oosterhuis makes a convincing attempt to do just that. In it he returns to the basics, or ‘goes forward to basics’ as he puts it, since he is dealing with basics that were not defined in such a way up to now. This theory of basics goes back further than comparable treatises that advance a collection of building components such as the column, the wall and the roof as the source of architecture. Instead, it is crucial here to think of a collection of points that display swarm behaviour.

Many theories on interactive buildings and environments are based on the idea of some sort of central ‘brain’, usually a computer controlled by an ‘all-knowing’ or ‘learning’ programme, that receives information from sensors and uses it to dictate behaviour. But a swarm has no central brain; a swarm consists of a great many ‘small, relatively stupid small brains’ that have learned to exchange information with their nearest neighbours. The behaviour of the collective is not directed centrally but stems from a large number of interactions among neighbours. Anyone who has ever seen a cloud of starlings will know that such a swarm can still display complex and seemingly ‘all knowing’ behaviour. There is no single leading starling that the whole flock follows. Rather, it is a case of a large group of starlings who are all both leaders and followers with respect to the other starlings flying close to them in the same swarm. That not only produces the most fantastic and organic configurations of the entire group, but also means that the swarm responds rapidly to changes in the surroundings, such as the presence of a predatory bird nearby. Programming such behaviour from a central brain would be an immensely difficult undertaking. But whoever thinks like starlings do, from the point (the individual flying starling) and his information exchange with his immediate neighbours, has things a lot easier as a programmer and benefits from the ‘wisdom of crowds’. As early as the 1980s Craig Reynolds succeeded in simulating the complex swarm behaviour excellently with a few programme rules, something that centrally thinking programmers still haven’t achieved.

That is why Oosterhuis proposes the building as a swarm of points that exchange information not only with their ‘nearest neighbours’, but also potentially with the surroundings, the climate and the people in and around the building. On the basis of that information exchange the ‘swarm’ (i.e. the building) can display behaviour. Such behaviour can take many forms. During the design process it can consist of searching for the optimal relation between structure and form in order to minimise the tensions in the structure. But it can also concern the interaction between usefulness, introduced through the functional information contained in the points, and the intuitive or formal ‘ideal of beauty’ of the designer. The information flow can be deployed to rationalise the production process. And after completion of the building this information exchange and interaction between points of intersection, climate and interior can continue and the building can ‘learn’ to respond to changes in the flow of information.

An architect who thinks in this way must cease to think primarily of the whole entity from now on. The point — for example a joint or a node in a structure, as is often the case with Kas Oosterhuis, but it could also be a plane or a member — is the basis of the idea. That point is controlled and programmed; and the behaviour of the points is what makes the whole entity (the swarm) adopt an unexpected form and display unexpected (but conscious nonetheless) behaviour. That does not mean that the designer no longer has anything to say about the form or behaviour of the whole but rather that, if he wants to change it, he must start with the point again, for example by adjusting the programme rules or by altering the flow of information to and/or from the points. The design process, and potentially the production and use process too, therefore involves constantly fine-tuning the nature and behaviour of the point in relation to the nature and behaviour of the whole entity.

‘Towards an New Kind of Building’ is written as a ‘designers guide for Nonstandard Architecture’, as the subtitle tells us, and is largely aimed at a professional audience. The book is divided into four chapters that cover successive stages towards a new way of building and designing. Interspersed among the more theoretical chapters are numerous anecdotes about and references to the daily practice of architecture by Oosterhuis and others. It is significant, for that matter, that the title first refers to building, and only after that to architecture. Building practice is the starting point; architecture is the result. If you understand and accept swarm behaviour as the basis, then the argumentation can follow easily.

The chapters repeatedly refer to the building as a body. Step 1 (chapter 1) is ‘Tag that Body’. All building components (points in the cloud/swarm) must be ‘tagged’ so that they can pass on information. That entails a lot of work during the early stages even though it doesn’t immediately produce visible results. But the benefits come in the later process.
Consequently, everything is much easier to manipulate in the second step, ‘Shape that body’.  That is the moment when the idea for the form, the aesthetics, can be introduced. It is also the moment when various experts such as engineers, climate advisors, production specialists — the architect is just one of many — can work on the model, if necessary at the same time. Or rather, they can add information to the points.
In Step 3 attention shifts from the design to the process of use. Whereas the behaviour took place in the computer during the design phase, it is of course obvious to continue that in the constructed reality. After all, why would a building stop reacting once it’s completed? If all its points have been programmed, then it is relatively simple to bring about interaction with the building’s occupants and the surroundings, even enabling the building to adapt through movement. Oosterhuis has already made several such interactive, moving spaces and structures with his Hyperbody group at the Faculty of Architecture at Delft University of Technology. Students turn out to be capable of programming and even constructing such a building relatively quickly — so it’s not that difficult.
The final step is ‘Evolve that Body’, in which Oosterhuis adds terms like personal emotion and evolution to the theory. In the end he proposes ‘buildings in real time’, as part of a much bigger entity, for example the ‘internet of things’, organisms with behaviour and with an autonomous evolution.

Whether we ever get that far remains the question, and even if we do it will take a generation or two. Just like the effects of the omnipresence of social media today, one could also critically question the future omnipresence of ‘things connected’ through the internet. In the meantime it is a fruitful exercise for everybody to ‘think forward from the basics’ with Kas Oosterhuis, even for architects who cling to the traditional top-down approach to architecture. Even if only to understand that there is another way. Who knows, maybe even they will be convinced that change is imperative.